Design, synthesis and trypanocidal activity of lead compounds based on inhibitors of parasite glycolysis

Article abstract of DOI:10.1016/j.bmc.2008.03.045

The glycolytic pathway has been considered a potential drug target against the parasitic protozoan species of Trypanosoma and Leishmania. We report the design and the synthesis of inhibitors targeted against Trypanosoma brucei phosphofructokinase (PFK) and Leishmania mexicana pyruvate kinase (PyK). Stepwise library synthesis and inhibitor design from a rational starting point identified furanose sugar amino amides as a novel class of inhibitors for both enzymes with IC₅₀ values of 23 μM and 26 μM against PFK and PyK, respectively. Trypanocidal activity also showed potency in the low micromolar range and confirms these inhibitors as promising candidates for the development towards the design of anti-trypanosomal drugs.

Full text of DOI:10.1016/j.bmc.2008.03.045

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 5050–5061
Design, synthesis and trypanocidal activity of lead compounds
based on inhibitors of parasite glycolysis
Matthew W. Nowicki,a,b Lindsay B. Tulloch,b Liam Worralll,b Iain W. McNae,b
c
c
b
´
Veronique Hannaert, Paul A. M. Michels, Linda A. Fothergill-Gilmore,
Malcolm D. Walkinshawb and Nicholas J. Turnera,*
aSchool of Chemistry, University of Edinburgh, King’s Buildings, West Mains Road, Edinburgh, EH9 3JJ, UK
bStructural Biochemistry Group, Institute of Structural and Molecular Biology, University of Edinburgh, King’s Buildings,
Mayfield Road, Edinburgh EH9 3JR, UK
cResearch Unit for Tropical Diseases, de Duve Institute and Laboratory of Biochemistry, Universite´ Catholique de Louvain (UCL),
Avenue Hippocrate 74, B-1200 Brussels, Belgium
Received 13 January 2008; revised 2 March 2008; accepted 11 March 2008
Available online 21 March 2008
Abstract—The glycolytic pathway has been considered a potential drug target against the parasitic protozoan species of Trypano-
soma and Leishmania. We report the design and the synthesis of inhibitors targeted against Trypanosoma brucei phosphofructoki-
nase (PFK) and Leishmania mexicana pyruvate kinase (PyK). Stepwise library synthesis and inhibitor design from a rational starting
point identified furanose sugar amino amides as a novel class of inhibitors for both enzymes with IC50 values of 23 lM and 26 lM
against PFK and PyK, respectively. Trypanocidal activity also showed potency in the low micromolar range and confirms these
inhibitors as promising candidates for the development towards the design of anti-trypanosomal drugs.
ꢀ 2008 Elsevier Ltd. All rights reserved.
1. Introduction
tential new drug targets in Trypanosomatidae.
Glycolysis appears to provide an excellent therapeutic
target because it is essential to bloodstream form Try-
panosomatidae as the only catabolic source of ATP.1,2
Furthermore, due to the evolutionary distance between
Trypanosomatidae and humans,3 the parasite enzymes
within the pathway possess distinct properties that dif-
ferentiate them from their mammalian counterparts,
and which could be exploited in the design of parasite-
specific drugs.4 The unique organisation of the glycolytic
pathway in trypanosomatids, with most of the enzymes
present in peroxisome-like organelles called glycosomes,
is correlated with exploitable differences. Thus, the com-
partmentalisation has resulted in different kinetic prop-
erties and activity regulation mechanisms that
correspond to differences in enzyme structure.4 Phos-
phofructokinase (PFK) and pyruvate kinase (PyK) are
validated as two such enzymes which could, therefore,
be targeted.2,4–6
The parasitic protozoa Trypanosoma and Leishmania
are the causative agents of the highly disabling and often
fatal diseases, such as African sleeping sickness, Chagas’
disease and leishmaniasis. Millions of people are at risk
in the areas of Africa, South America and Asia where
these diseases are endemic. Unfortunately, current drugs
are far from satisfactory as they often possess toxic side
effects, are ineffective against certain disease forms and
may require expensive administration procedures; in
addition, resistance to these drugs is becoming an
increasingly common problem.1
In the last two decades, experimental approaches such
as gene cloning, protein crystallography and RNA inter-
ference have been introduced to define and validate po-
Keywords: Pyruvate kinase; Phosphofructokinase; Inhibitors; Leish-
mania; Trypanosome; Parasite; Drug development; Glycolysis.
PFK catalyses the phosphorylation of fructose 6-phos-
phate (F6P) to produce fructose 1,6-bisphosphate
(F1,6BP). The parasite enzyme exists as a homotetramer
and, under physiological conditions, the reaction is
*
Corresponding author at present address: School of Chemistry,
Manchester Interdisciplinary Biocentre, 131 Princess Street, Man-
chester M1 7DN, UK; e-mail: Nicholas.Turner@Manchester.ac.uk
0968-0896/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.03.045

M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
5051
essentially irreversible. Recently, the crystal structure of
Trypanosoma brucei PFK (Tb PFK) has been solved7
showing striking differences from human PFK. Not only
is there a low sequence identity between the two enzymes
(<25% identical), but the Tb PFK structure shows an
additional domain and loops, including a loop adjacent
to the active site which is not present in the human struc-
ture. Such differences are important in being able to de-
sign selective inhibitors of the parasitic PFK targeted
towards the active site. Indeed, preliminary attempts
to design inhibitors of trypanosomatid PFK have had
promising results. Analogues of 2,5-anhydro-D-manni-
tol bearing various arylamino groups at position 1
(structurally similar to F1,6BP, Scheme 1) have shown
weak inhibitory activity by reversibly binding at the
ATP and F6P binding sites,8 the most active of which
bears a phosphate at the 6-position. Although in most
eukaryotes, including humans, fructose 2,6-bisphos-
phate (F2,6BP) is a potent activator of PFK, this is
not the case in trypanosomatids. Thus, fructose–phos-
phate analogues are expected to affect trypanosomatid
PFKs through the active site, not the effector site.
scaffold (series I, Scheme 1) primarily by substitution at
the 1-position. These compounds were screened against
both Tb PFK and Lm PyK (due to the structural similar-
ity between F6P, F2,6BP and the inhibitor scaffold). A
further round of inhibitor development involved func-
tionalising the 6-position through selective oxidation,
followed by amide coupling.
The results that are reported herein are an important
step towards the overall goal of producing trypanocidal
drugs. Due to the high species similarities of PFKs and
of PyKs within the Trypanosomatidae (>70% sequence
identity for each enzyme), it is likely that inhibitors that
are active against one parasite, would also be efficacious
against Trypanosoma cruzi and the various pathogenic
Leishmania species.
2. Chemistry
Derivatives in series I (Scheme 1) were prepared in two
steps following the procedures described by Claustre
et al.8,10 The nitrous acid deamination of D-(+)-glucosa-
mine hydrochloride, an adaptation of the Horton and
Philips method,11 proceeded in good yield and was fol-
lowed by reductive amination reactions using various
primary amines, adapted for multi-parallel synthesis
(low to moderate yields, purified using mass-directed
purification). Amines in series I were screened against
Tb PFK and Lm PyK.
PyK catalyses the conversion of phosphoenolpyruvate
(PEP) to pyruvate whilst producing ATP. Trypanoso-
matid PyK is allosterically regulated by F2,6BP which
induces conversion to the active R-state of the enzyme
from the inactive T-state.6 In contrast, mammals possess
four isoenzymes of PyK, one of which (M1) is constitu-
tively active and the other three (M2, L and R) are reg-
ulated by F1,6BP.4 Although Tb PyK can be activated
by F1,6BP, the concentrations needed are 2000-fold
higher than those needed for F2,6BP. The determination
of the crystal structure of Leishmania mexicana PyK
(Lm Pyk)9 and comparisons with human PyKs have
highlighted significant differences at the effector site that
would, therefore, appear to be another promising target.
Modification of the reductive amination step was nec-
essary in order to improve yields for subsequent
inhibitor development and the generation of series
II; this improvement was achieved using a solid sup-
ported borohydride resin and the amine scavenger re-
sin, 4-benzoxybenzaldehyde polystyrene.12–14 The
mannonamide derivatives in series II were prepared
in four steps from amines 2m and 3m. Boc protection
of the secondary amine occurred in moderate yields,
followed by the key TEMPO mediated selective oxi-
The starting point for inhibitor development was the
synthesis of a wide range of compounds based on the
N-substituted-1-amino-2,5-anhydro-1-deoxy-D-mannitol
Scheme 1. Reagents and conditions: (a) NaNO2, Amberlite 120 H+, 0 ꢁC, 18 h; (b) Dowex CO 2ꢀ; (c) RNH2, (20 min for parallel synthesis, 2 h when
3
using solid phase reagents); (d) NaBH3CN, HCl, 1.5 h; (e) borohydride resin, 24 h; (f) 4-benzoxybenzaldehyde polystyrene, 72 h; (g) Boc2O, DMAP,
dioxane, MeCN, H2O, 0 ꢁC, 2 h; (h) TEMPO, KBr, TBACl, NaOCl, NaCO3, brine, DCM, 0–25 ꢁC, 2–24 h; (i) R0NH2, EDCI, HOBt, DIPEA,
DCM, 20 h; (j) TFA, DCM, 2 h.

5052
M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
dation of the 6-hydroxyl to the corresponding acid.
The biphasic oxidation was an adaptation of the
method set out by Davis and Flitsch15 and gave
low to moderate yields, although this could have
been due to the difficulty in extracting the target
acids from the aqueous phase. Subsequent amide cou-
pling using EDCI, HOBt and DIPEA16,17 and Boc
deprotection gave the series II sugar amino amides
which were screened against either Tb PFK, Lm
PyK or both, depending upon the R group specificity
from the series I screen.
low millimolar range), however, amine 3m gave an
IC50 of 71 lM which is over 10-fold more potent than
any other inhibitors from the same series. This difference
is striking and points to the suggestion that there is a
hydrophobic pocket located in the effector site which al-
lows for a highly specific interaction with the cycloheptyl
moiety of 3m. Of the other amines in series I that bear a
carbocycle at the 1-position, all showed much lower po-
tency than 3m indicating that the hydrophobic interac-
tion is likely to be specific. Moreover, amine 3m is a
novel inhibitor of Lm PyK and was an ideal candidate
to take forward for the development of series II.
3. Results and discussion
3.1. Library 1
3.2. Library 2
Table 2 shows inhibitory activity of compounds in series
II against both Tb PFK and Lm PyK. All members of
series II are sugar amino amides, derived from the two
best hits (2m and 3m) from the series I screen.
The discovery that analogues derived from 2,5-anhydro-
D-mannitol showed weak inhibition against Tb PFK8
led to further investigation into inhibitors of this type
(series I). Series I was additionally screened against
Lm PyK because F1,6BP is not only the product of
PFK, but is also an effector of PyK.
Activity results against Tb PFK showed a noticeable
improvement over series I. The initial screen was per-
formed at an inhibitor concentration of 1 mM (in con-
trast to the 5 mM concentration used for series I) and
all amides tested exhibited an improved level of inhibi-
tion over amine 2m. Determined IC50 values for four
of the best inhibitors demonstrate around a 10-fold in-
crease in potency compared to amine 2m and to previ-
ous inhibitors of Tb PFK8 confirming that the sugar
amino amides in series II are a new class of Tb PFK
inhibitor. Amide 8e exhibited the greatest inhibition
with an IC50 of 23 lM although other amides displayed
comparable potency (low micromolar) showing that the
introduction of the amide group has a marked effect
upon potency of the series II analogues.
Inhibitory activities of 55 amine derivatives of 2,5-
anhydro-D-mannitol in series I against Tb PFK are re-
ported in Table 1. All compounds (tested at 5 mM)
showed inhibition to some extent, with 13 compounds
causing greater than 50% inhibition. Compound 2m
(2,5-anhydro-1-deoxy-(3,4-dichlorobenzylamino)-D-man-
nitol) was particularly effective by reducing activity to
5%. Compound 2m therefore possessed at least a
6-fold better potency than other amines in this screen.
The IC50 for 2m was determined to be 0.41 0.05
mM, a value comparable to the most potent analogue
described by Claustre et al.8 (R = 3-nitrophenyl,
IC50 = 0.45 mM). However, it should be noted that
the 3-nitrophenyl derivative is phosphorylated at the
6-position, and when non-phosphorylated had an al-
most doubled IC50 value of 1.1 mM. The results from
series I thus identified an analogue with a greatly
optimised R-group (3,4-dichlorobenzyl) which was
incorporated in further inhibitor development leading
to the synthesis of series II.
Figure 1a depicts a representation of potential interac-
tions of inhibitor 8e with the F6P binding site of Tb
PFK (pdb id: 2HIG). The model has been generated
by positioning the furanose ring of 8e in the position
in which F1,6BP is found in the E. coli PFK structure
(1PFK)18 so that the 3,4-trans-dihydroxyl moiety is
clamped by residue Glu325. In the orientation shown,
the amide is making three hydrogen bonding interac-
tions with Arg173 and Arg383, which inhibitor 2m
would not be able to make. Thus, coupled with the extra
hydrophobic bulk of the series II inhibitors, the interac-
tions they gain could account for their increase in po-
tency over the series I inhibitors. Interestingly, it can
be seen from the results in Table 2 that when the 3,4-
dichlorobenzyl group and the R0 group are reversed
(as in the case for amides 8f and 8h) little change in po-
tency is observed (IC50 = 80 lM for 8f and 49 lM for
8h). An explanation for this result could be that either
the hydrogen bonding interactions gained by the pres-
ence of the amide group compensate for the switching
of the optimised 3,4-dichlorobenzyl group, or that the
sugar amino amide inhibitors have a different binding
mode compared to that of 2m. Due to the pseudo-C2
symmetry of the furanose scaffold of the series II inhib-
itors, it is likely that the 3,4-dichlorobenzyl group is
actually the optimal group for the 6-position, rendering
the 1-position susceptible to further optimisation. A
Screening of series I against Lm PyK gave 16 com-
pounds with greater than 50% inhibition (Table 1), with
eight compounds reducing their activity to 30% or less.
Interestingly, it appears that the most active of these
bear hydrophobic groups at the 1-position. It is also
noticeable that whilst the majority of series I compounds
inhibit PyK, there are a small number of compounds
that activate PyK (remaining activity greater than
100%), which would suggest that the series may bind
to the effector site rather than binding to the active site
of the protein. (Compounds that bind to the active site
would only exhibit inhibition of the enzyme, whilst those
that bind to the effector site could exhibit either inhibi-
tion or activation.)
IC50 values of ten of the more active compounds were
determined and are shown in Table 1. The majority of
the hits showed only weak inhibitory activity (in the

M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
Table 1. Screening results of series I against Tb PFK and Lm PyK
5053
Compound
R
PFK screen (% r.a.)a
58
IC50 vs PFK (mM)b
PyK screen (% r.a.)a
—
IC50s vs PyK (mM)b
2a
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
2b
2c
2d
2e
2f
85
65
65
73
61
68
30
46
63
59
57
5
—
—
—
—
—
—
—
—
—
—
—
0.41
—
—
—
—
—
88
81
71
90
89
2g
2h
2i
100
89
105
42
2j
2k
2l
107
100
43
2m
2n
2o
2p
2q
2r
69
87
59
75
87
—
103
85
—
52
4.1
(continued on next page)

5054
M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
Table 1 (continued)
Compound
R
PFK screen (% r.a.)a
IC50 vs PFK (mM)b
PyK screen (% r.a.)a
IC50s vs PyK (mM)b
2s
2t
83
70
—
—
72
26
—
2.5
2u
2v
2w
2x
2y
2z
64
77
33
57
52
62
—
—
—
—
—
—
80
97
—
—
—
—
—
1.2
49
60
103
36
3a
3b
54
—
—
—
43
40
4.3
—
3c
3d
3e
3f
—
—
—
—
—
—
69
—
33
58
63
—
—
—
—
—
—
—
—
—
—
—
22
25
36
53
58
84
83
30
34
125
19
3.5
0.9
—
—
3g
3h
3i
—
—
—
3j
1.5
1.5
—
3k
3l
3m
0.071

M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
5055
Table 1 (continued)
Compound
R
PFK screen (% r.a.)a
64
IC50 vs PFK (mM)b
PyK screen (% r.a.)a
57
IC50s vs PyK (mM)b
3n
3o
3p
3q
3r
—
—
2.3
—
—
—
—
—
1.5
—
—
—
—
—
—
—
—
49
51
44
87
—
—
40
73
64
52
35
36
41
47
47
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
26
86
115
80
28
47
19
95
73
86
82
90
86
85
59
3s
3t
3u
3v
3w
3x
3y
3z
4a
4b
4c
Screenings were performed at an inhibitor concentration of 5 mM; r.a. = remaining activity.
a Screenings performed in triplicate where sufficient compound was available. In general, SD <5%.
b R2 values >0.99.
proposed extension of this library could, therefore,
include sugar amino amides where the 1-position is
modified in the presence of the 6-position 3,4-dichlorob-
enzylamido group.
amide group does not appear to have as great an effect
upon inhibition although a high proportion of hits are
observed, signifying the importance of the cycloheptyl
group on potency. Almost a 3-fold improvement in po-
tency is seen over the best hit from series I, with 8m hav-
ing an IC50 value of 26 lM. In the case where the
optimised cycloheptyl group and the R0 group are re-
versed (as for compounds 8m and 8f), all potency is lost
Results of the screen of series II against Lm PyK showed
comparable results to the best hits from the series I
screen. Unlike for Tb PFK, the introduction of the

5056
M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
Table 2. Screening results of series II against Tb PFK and Lm PyK
Compound General Structure R0
PFK screen (% r.a.)a IC50 vs PFK (lM)b PyK screen (% r.a.)a IC50 vs PyK (lM)b
8a
8b
8c
8d
8e
8f
A
A
A
A
A
A
A
B
B
B
B
B
B
B
4.0
2.0
1.7
0.8
0.9
4.7
5.5
0.4
—
—
50
—
—
23
80
—
49
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
53
—
26
—
—
—
—
—
88.2
—
8g
8h
8i
—
11.5
19.1
7.6
12.5
1.9
24.8
8j
—
8k
8l
—
—
8m
—
8n
—
Inhibitor concentration was 1 mM for PFK screen and 5 mM for PyK screen; r.a. = remaining activity.
a Screenings performed in triplicate where sufficient compound was available. In general, SD <5%.
b R2 values >0.99.
(only 12% inhibition at a concentration of 5 mM) clearly
indicating that the 1-position is optimised for the cyclo-
heptyl group, and that the hydrophobic pocket identi-
fied in the series I screen cannot accommodate an
amide functionality (likely to be caused by electrostatic
repulsion of the amide carbonyl).
model the furanose ring of 8m in the same orientation
as F1,6BP in the yeast PyK structure (pdb id: 1A3W).
However, in this position 8m is unable to occupy the
effector site due to severe steric clashes. It should be
noted that the position of the flexible loop in the yeast
structure (R-state) is in a more open conformation com-
pared to that in the Lm structure (T-state). However,
steric clashes are not just caused by the flexible loop,
but are also caused by other clashes with both cyclohep-
Figure 1b shows a model of 8m in the effector site of Lm
PyK (pdb id: 1PKL). Initially, attempts were made to

M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
5057
inhibit the necessary activating conformational changes.
The increased potency observed between inhibitor 3m
and the series II inhibitors could be largely due to the in-
crease in hydrophobic bulk. Therefore, 8m is a good
starting point for lead optimisation, particularly for
alternative modifications at the 6-position.
3.3. In vitro parasite screens
An important part of current drug design is to verify the
inhibitor’s action on the biological system for which it
was designed. Even at such an early stage in the drug de-
sign process, evidence of activity against target cells is
desirable to help with subsequent lead identification
and modification as well as to confirm the desired bio-
logical effect. Information gained may also help with
identifying inhibitors that have the correct physical
properties to enter the target cells as these properties
could play a crucial role in further rounds of inhibitor
modification. Table 3 shows the results of some series
I and II inhibitors tested against cultured T. brucei
brucei cells (strain 427).
Significantly, the in vitro screening results show that all
inhibitors tested show some form of activity against the
cultured parasite, although the series II inhibitors are
rather more potent than series I, the most potent being
8e with an ED50 of 30 lM. This is a crucial and very
important result that not only reinforces the inhibitor
design strategy but also confirms the direction for lead
modification. Furthermore, it confirms that the glyco-
lytic pathway is an excellent target for drug design, as
even at an early stage of inhibitor identification and
modification, a reasonable level of activity is observed
against the target parasites.
Figure 1. Models of potential interactions of (a) 8e with the active site
of Tb PFK, and (b) 8m with the effector site of Lm PyK.
4. Conclusion
Currently, there is a great need to find new drugs and
remedies to combat the growing incidence of parasitic
diseases that threaten many African, Asian, Central
and Southern American countries.1 From a weakly
binding inhibitor (for Tb PFK) as a starting point,8
through library design and synthesis, new compounds
were identified for both Tb PFK and Lm PyK. A fur-
ther round of inhibitor development identified fura-
tyl moieties of 8m. Therefore, a simple overlay of the
furanose rings of F1,6BP and 8m seems unlikely. An
alternative mode of binding was discovered by rotating
the sugar scaffold through 180 degrees. The cyclohep-
tylamino moiety now occupies a hydrophobic pocket
created by residues Leu399, Val457, Val478 and the C4
chain of Lys453, with the furanose scaffold still making
interactions with the flexible loop and possibly the side
chain of Lys453. The mechanism for allosteric control
of Lm PyK is not fully understood, although it is likely
to arise via conformational changes in the effector site
(movement of the ‘flexible’ and ‘variable’ loops, Figure
1b) which cause further conformational changes to fa-
vour binding of the substrate in the active site.9,19 Our
docked model (Fig. 1b) suggests that the inhibitors of
Lm PyK may bind to and stabilise the inactive form of
the enzyme, and with the cycloheptyl moiety anchored
in the hydrophobic pocket, residues Lys453 and
His480 are unable to interact with an electronegative
group (such as the 2-phosphate of F2,6BP) and, there-
fore, the transition from T to R state is restricted. The
hydrophobic interactions themselves are also likely to
nose sugar amino amides as
a
new class of
trypanosomal glycolytic pathway inhibitors, with 8e
being the lead against PFK (IC50 = 23 lM) and 8m
the lead against PyK (IC50 = 26 lM). Modelling stud-
Table 3. Results of in vitro inhibitor screenings against Trypanosoma
brucei
Inhibitor
ED50 (mM)
2m
2z
0.13
0.83
0.40
1.88
0.11
0.030
0.035
3b
3m
8b
8e
8f

5058
M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
ies on known crystal structures of Tb PFK and Lm
PyK have provided some insight into the possible
modes of action of these inhibitors and will help guide
lead development and further probing of the respec-
tive binding sites. In the case of PyK, effector site
probing with specifically designed inhibitors may also
provide information as to its allosteric mechanism.
Crucially, the effectiveness of the inhibitors against
cultured trypanosomatids has not only shown a good
level of trypanocidal activity, but confirms the sugar
amino amides as prime candidates for lead
development.
moved by filtration and the remaining solution was
lyophilised to yield the title compound as a pale yel-
low hygroscopic solid (1.33 g, 88%). Rf = 0.42,
CH2Cl2/MeOH (8:2) 1H NMR (D2O),
d (ppm):
3.50–3.75 (3H, m, H-2, 2· H-6), 3.78–3.86 (1H, m,
H-5), 3.95 (1H, t, J4,3 = J4,5 = 5.5 Hz, H-4), 4.10 (1H,
t, J3,4 = J3,2 = 5.5 Hz, H-3), 5.00 (1H, d, J1,2
=
5.5 Hz, H-1 gem-diol), 8.35 (s, H-1 aldehyde). 13C
NMR (D2O), d (ppm): 61.2 (C6), 76.8 (C4), 77.8
(C3), 83.2 (C5), 84.6 (C2), 90.0 (C-1 gem-diol).
5.1.2. Synthesis of series I: N-substituted-1-amino-2,5-
anhydro-1-deoxy-D-mannitol derivatives (2a–4c).20 Using
the Bohdan Neptune reagent manager, a methanolic
solution of 1 (0.7 M) was added to 17· 2 mL reactors
each containing a benzylamine derivative (100 mg, pre-
weighed by Sigma–Aldrich Filling Station) in equimolar
quantities. The Bohdan MiniBlock synthesiser that con-
tained the reactors was then shaken for 15 min. Sodium
cyanoborohydride in methanol (2.4 M, 0.7 equiv) was
added to each reactor, before adjusting the pH to 6 by
dropwise addition of HCl in methanol (0.5 M). The
MiniBlock was then shaken for a further 2 h, monitor-
ing the pH of each reactor, and adjusting the pH if nec-
essary. The reaction mixtures were then transferred to a
microtitre plate and the solvent removed under reduced
pressure. The residues were dissolved in acetonitrile/
water (1:1, 0.1% acetic acid) and purified using mass di-
rected purification.21
5. Experimental
5.1. Materials and methods
1H and 13C NMR were recorded on Bruker AC 250 or
AC 360 instruments. Chemical shifts (d) were recorded
in parts per million (ppm) downfield of tetramethylsil-
ane and were referred to residual undeuterated solvent
present in the deuterated solvent of the sample (e.g.,
CH3OH in CD3OD). Electrospray ionisation mass spec-
troscopy (ESI) was carried out on a Micromass VG
Platform II instrument. Solvents used were either water
or acetonitrile/water (1:1, 0.1% acetic acid) in either po-
sitive or negative scan modes. Fast Atom Bombardment
(FAB) high resolution mass spectra were recorded on a
Kratos MS50TC instrument. Infra-red absorption spec-
troscopy was performed on a Perkin-Elmer FT-IR Par-
agon 1000 spectrometer, and mmax values are quoted in
cmꢀ1. Optical rotations were performed at room tem-
perature on an AA1000 polarimeter from Optical Activ-
ity Ltd (measurements made at the sodium D-line).
Concentrations are given in g/100 mL. Thin layer chro-
matography was performed on Merck DC-Alufolien
Kieselgel 60F245 0.2 mm pre-coated plates. Components
were visualised by UV fluorescence or by potassium
permanganate or ninhydrin stains. Melting points were
obtained on Gallenkamp melting point apparatus. Par-
allel synthesis was performed using the following instru-
ments: Bohdan Neptune reagent dispenser, Bohdan
MiniBlock Synthesiser and shaker, and Christ b-RVC
solvent evaporator. Mass directed purification was car-
ried out on a Waters ZMD mass spectrometer with
Waters 600 HPLC using MassLynx version 3.5 and
FractionLynx software. The HPLC column used was a
Waters Xterra, C18, 60 · 21.20 mm, 5 u microns. Paral-
lel flash column chromatography was carried out using
Quad3 parallel purification system using pre-packed
columns.
5.1.3. 2,5-Anhydro-1-deoxy-1-(3,4-dichlorobenzylamino)-
D-mannitol (2m). From 3,4-dichlorobenzylamine and
using ZMD method 1, yield = 12 mg, 7%. Decomposes
1
>190 ꢁC. H NMR (CD3OD), d (ppm): 2.72 (1H, dd,
Jgem = 12.5 Hz, J1a,2 = 7.5 Hz, H-1a), 2.79 (1H, dd,
Jgem = 12.5 Hz, J1b,2 = 3.5 Hz, H-1b), 3.61–4.03 (8H,
m, H-2, H-3, H-4, H-5, 2· H-6, 2· ArCH2), 7.24 (1H,
dd, JArH-ortho = 8.0 Hz, JArH-meta = 2.0 Hz, ArH),
7.40 (1H, d, JArH-ortho = 8.0 Hz, ArH), 7.58 (1H, d,
JArH-meta = 2.0 Hz, ArH). 13C NMR (CD3OD),
d
(ppm): 50.1 (C1), 51.3 (Ar-CH2), 61.5 (C6), 76.9 (C2),
79.2 (C5), 81.7 (C4), 83.5 (C3), 127.4–129.6 (3C, 3·
ArH), 129.9–131.3 (2C, 2· ArCl), 139.9 (C8, ArC).
ESI+: [M+H]+ = 322.1 [35Cl–35Cl]. FAB HRMS (+ve)
found m/z = 322.06074 and 324.05862 [M+H]+,
C13H18NO4Cl2 requires 322.06129 (2· 35Cl) and
324.05834 (35Cl + 37Cl). [a]D = +5.7ꢁ (c 0.7, MeOH).
5.1.4. 2,5-Anhydro-1-deoxy-1-(cycloheptylamino)-D-man-
nitol (3m). From cycloheptylamine and using ZMD
Method 2, yield = 66 mg, 33%. Decomposes >225 ꢁC.
1H NMR (CD3OD), d (ppm): 1.27–1.91 (12H, m, 6·
–CH2-cycloheptane), 2.65–2.84 (3H, m, 2· H-1, CH
5.1.1. 2,5-Anhydro-D-mannose (1).10 D-Glucosamine
hydrochloride (2.0 g) was dissolved in water (50 mL)
and stirred at 0 ꢁC for approximately 5 h. Sodium ni-
trite (1.6 g, 2.5 equiv) was then added, followed by
cautious addition of Amberlite 120 H+ resin
(47 mL), maintaining the temperature between 0 and
5 ꢁC. The reaction was then left to stir at 0 ꢁC for
18 h before the resin was removed by filtration. The
cycloheptane),
3.60
(1H,
dd,
Jgem = 12.0 Hz,
J5,6a = 5.5 Hz, H-6a), 3.68 (1H, dd, Jgem = 12.0 Hz,
J5,6b = 3.5 Hz, H-6b), 3.78–3.93 (4H, m, H-2, H-3, H-
4, H-5). 13C NMR (CD3OD), d (ppm): 23.6 (2C, 2·
cycloheptane), 27.3 (2C, 2· cycloheptane), 33.0 (2C, 2·
cycoheptane), 48.4 (C1), 58.6 (cycloheptane CH), 61.5
(C6), 77.0 (C4), 79.5 (C3), 81.6 (C2), 84.1 (C5). ESI+:
[M+H]+ = 260.2. FABHRMS (+ve) found m/
z = 260.18594 [M+H]+, C13H26NO4 requires 260.18618.
[a]D = 6.7ꢁ (c 1.7, MeOH).
solution was then neutralised by portion-by-portion
2ꢀ
addition of Dowex CO3
resin. The resin was re-

M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
5059
5.1.5. Scale-up method for amines 2m and 3m. The
appropriate amine (1.5 equiv) was added to a solution
of aldehyde 1 (0.36 g, 1 equiv) in methanol (10 mL) in
a solid phase reaction cartridge. The cartridge was
sealed and placed on a shaker for 2 h. Borohydride resin
(1.17 g, 2 equiv, 3.8 mmol/g) was added and the car-
tridge was shaken for a further 24 h. 4-Benzoxybenzal-
dehyde polystyrene (0.76 g, 1.1 equiv, 3.2 mmol/g) in
dichloromethane (4 mL) was added before leaving the
reaction to shake for a further 72 h. After ESMS mass
spectroscopy (positive) confirmed no more 3,4-dichloro-
benzylamine to be present in the solution, the resin was
removed by filtration and was washed with methanol
(3· 5 mL) and dichloromethane (5 mL). All filtrates
were combined and the solvent removed under reduced
pressure to yield the title compounds (yield = 84–94%).
5.1.8. N-Boc-2,5-anhydro-1-deoxy-1-(3,4-dichlorobenzy-
lamino)-D-mannonic acid (6a). N-Boc protected amine
5a (0.13 g, 1 equiv) and TEMPO (3 mg, 6 mol%) were
dissolved in CH2Cl2 (1 mL). A solution of KBr
(0.4 M), tetrabutylammonium chloride (0.3 M) in sat.
NaHCO3 (1 mL) was added and the solution was cooled
to 0 ꢁC and stirred vigorously. A solution of sodium
hypochlorite (1.3 M, 0.75 mL), sat. NaHCO3 (0.33 mL)
and brine (0.66 mL) were added dropwise over a period
of 30 min. The reaction was then allowed to stir for a fur-
ther 2 h. The two layers were then separated and the or-
ganic phase was washed with water (3· 10 mL). The
aqueous layers were combined and acidified to pH 1 with
HCl (2 M) before extracting with ethyl acetate (3·
10 mL). The organic layer was then dried over Na2SO4,
filtered and the solvent removed under reduced pressure
to yield the title compound as an off white solid (97 mg,
72%). Decomposes >150 ꢁC. FTIR thin film (cmꢀ1):
3417 br (OH), 2978 and 2932 (CH), 1667 (2· C@O).
1H NMR (CDCl3), d (ppm): 1.47 (9H, s, 9· tBu CH3),
3.24–3.47 (2H, m, 2· H-1), 3.84–4.40 (6H, m, H-2, H-
3, H-4, H-5, 2· ArCH2), 5.93 (1H, s broad, acid OH),
6.98 (1H, d, JArH-ortho = 7.5 Hz, ArH), 7.24 (1H, s,
ArH), 7.31 (1H, d, JArH-ortho = 7.5 Hz, ArH). 13C
NMR (CDCl3), d (ppm): 28.2 (3· tBu CH3), 48.0 (C1),
51.0 (ArCH2), 78.0 (C4), 79.7 (C3), 81.5 (tBu C), 81.8
(C5), 83.9 (C2), 126.5–130.3 (3C, 3· ArH), 131.0–132.3
(2C, 2· ArCl), 138.3 (ArC), 156.4 (Boc carbonyl),
174.1 (acid). ESIꢀ: [M–H]ꢀ = 434.1 [35Cl–35Cl].
5.1.6. N-Boc-2,5-anhydro-1-deoxy-1-(3,4-dichlorobenzy-
lamino)-D-mannitol (5a). To a stirred mixture of amine
2m (0.96 g, 1 equiv) in acetonitrile (10 mL), water was
added dropwise until the amine had dissolved. Di-tert-
butyl dicarbonate (0.64 g, 1 equiv) in dioxane (2 mL)
was added and the reaction was cooled to 0 ꢁC. 4-
Dimethylaminopyridine (0.091 g, 0.25 equiv) was added
then and the reaction was stirred for 2 h. After the addi-
tion of water (10 mL), the reaction was extracted in to
ethyl acetate (3· 10 mL). The organic layer was washed
with sat. NaHCO3 (2· 10 mL) and the combined aque-
ous layers were neutralised with 2 M HCl before being
re-extracted with ethyl acetate (2· 10 mL). The com-
bined aqueous layers were then acidified to pH 1 before
another re-extraction with ethyl acetate (3· 10 mL). All
organic layers were combined and washed with 5%
NaHSO4 (2· 5 mL), dried over Na2SO4, filtered and
the solvent removed under reduced pressure to yield
the title compound as a pale yellow oil (0.73 g, 58%).
5.1.9. N-Boc-2,5-anhydro-1-deoxy-1-(cycloheptylamino)-
D-mannonic acid (6b). Prepared from 5b using the same
method as for the preparation of 6a with the stir time in-
1
creased from 2 to 24 h (yield = 0.19 g, 38%). H NMR
(CDCl3), d (ppm): 1.38–1.74 (21H, m, 6· –CH2-cyclo-
heptane, 9· tBu CH3), 3.18–3.41 (2H, m, 2· H-1),
3.42–5.06 (5H, m, H-2, H-3, H-4, H-5, cycloheptane
CH). ESIꢀ: [M–H]ꢀ = 372.1.
1
Rf = 0.27 CH2Cl2/MeOH (9.5:0.5). H NMR (CDCl3),
d (ppm): 1.34 (9H, s, 9· tBu CH3), 3.24–3.46 (2H, m,
2· H-1), 3.59–3.73 (2H, m, 2· H-6), 3.75–4.14 (4H, m,
H-2, H-3, H-4, H-5), 4.37 (2H, s, 2· ArCH2), 6.98
(1H, dd, JArH-ortho = 8.0 Hz, JArH-meta = 2.0 Hz, ArH),
7.24 (1H, d, JArH-meta = 2.0 Hz, ArH), 7.32 (1H, d,
5.1.10. Synthesis series II: N,N0-substituted-1-amino-2,5-
anhydro-1-deoxy-1-D-mannonamide derivatives. Part 1:
amine coupling (7a–7n).22 The N-Boc protected sugar
amino acids 6a and 6b (1 equiv) were dissolved in CH2Cl2
(1 mL) and placed into 10 mL reaction tubes (6· 90, 7·
91). To each tube a different amine was added (1 equiv,
7 amines used in total) along with EDCI (1.3 equiv),
HOBt (1.5 equiv) and DIPEA (2.2 equiv). The reactions
were allowed to stir for 20 h before quenching with NaCl
(2 M, 1 mL). For each reaction, the two phases were sep-
arated and the aqueous phase extracted with CH2CH2 (2·
2 mL). The organic phases were combined and washed
with water (2· 2 mL) and HCl (0.5 M, 2· 2 mL). The solu-
tions were then dried over Na2SO4, filtered and the sol-
vent removed under reduced pressure to yield the crude
products. The compounds were purified using the Quad3
parallel purification system, using pre-packed columns,
eluting with CH2Cl2/MeOH (9.5:0.5). Solvent removal
yielded the desired compounds.
JArH-ortho = 8.0 Hz, ArH). 13C NMR (CDCl3),
d
(ppm): 28.2 (3C, 3· tBu CH3), 48.2 (C2), 51.0 (ArCH2),
62.1 (C6), 76.6 (C4), 78.5 (C3), 81.2 (tBu C), 82.3 (C5),
82.9 (C2), 126.4–130.3 (3C, 3· ArH), 131.0–132.3 (2C,
2· ArCl), 132.3 (ArC), 156.6 (Boc carbonyl). ESI+:
[M+Na]+ = 444.2 [35Cl–35Cl]. FTIR thin film (cmꢀ1):
3398 br (OH), 2978 and 2932 (CH), 1672 (C@O).
5.1.7. N-Boc-2,5-anhydro-1-deoxy-1-(cycloheptylamino)-
D-mannitol (5b). Prepared from 3m using the same
method as for the preparation of 5a (yield = 0.88 g,
57%). 1H NMR (CDCl3), d (ppm): 1.37–1.78 (21H,
m, 6· –CH2-cycloheptane, 9· tBu CH3), 3.05–3.39
(2H, m, 2· H-1), 3.64–4.32 (7H, m, H-2, H-3, H-4,
H-5, 2· H-6, CH cycloheptane). 13C NMR (CDCl3),
d (ppm): 23.5–27.8 (4C, 4· cycloheptane), 28.3 (3C,
3· tBu CH3), 33.2–35.3 (2C, 2· cycloheptane), 47.2
(C1), 60.9 (cycloheptane CH), 62.6 (C6), 77.0
(C4), 78.4 (C3), 80.4 (tBu C), 83.4 (C5), 83.6 (C2),
5.1.11. N-Boc-2,5-Anhydro-1-deoxy-1-(3,4-dichloroben-
zylamino)-D-3,4-dichlorobenzylmannonamide (7e). From
3,4-dichlorobenzylamine and 6a, yield = 33 mg, 25%.
156.9
382.1.
(Boc
carbonyl).
ESI+:
[M+Na]+ =
1
Rf = 0.14 CH2Cl2/MeOH (9.5:0.5). H NMR (CDCl3),

5060
M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
d (ppm): 1.39 (9H, s, 9· tBu CH3), 3.28–3.48 (2H, m, 2·
H-1), 3.81–4.41 (8H, m, H-2, H-3, H-4, H-5, 4· ArCH2),
6.99 (2H, 2· ArH), 7.20–7.32 (4H, m, 4· ArH).
the coupling enzymes alone to avoid a false positive re-
sult. All compounds were found to inhibit PyK only.
5.1.16. In vivo screens against Trypanosoma brucei. The
anti-trypanosomal activity tests were performed accord-
ing to the procedure previously reported by Hoet et al.24
using cultured bloodstream-form cells of strain T. brucei
427. A 76-h incubation time was selected.
5.1.12. Synthesis of series II: Synthesis of N,N0-substi-
tuted-1-amino-2,5-anhydro-1-deoxy-1-D-mannonamide deriv-
atives. Part 2: Boc-deprotection (8a–8n).23 Boc-protected
amino mannonamides (from part 1) were dissolved in
CH2Cl2 (1.5 mL). Trifluoroacetic acid (1 mL) was added
and the reactions were stirred for 2 h. The solvent was
then removed under reduced pressure and the residue
dissolved in acetonitrile/water (1:1, 0.1% acetic acid)
and purified using mass directed purification (ZMD
Method 1). The removal of the solvent by lyophilisation
afforded the desired compounds.
Acknowledgments
We thank various labworkers and honours students who
helped during the course of this project, particularly
´
Jose Martinez-Oyanedel and Marcos Navarro. This
work was supported by the European Commission un-
der its INCO-DEV programme.
5.1.13. 2,5-Anhydro-1-deoxy-1-(3,4-dichlorobenzylamino)-
D-3,4-dichlorobenzylmannonamide (8e). From 7e, yield =
16 mg, 58%. 1H NMR (CD3OD), d (ppm): 3.28 (1H, dd,
Jgem = 13.0 Hz, J1a,2 = 3.0 Hz, H-1a), 3.46 (1H, dd,
Supplementary data
Jgem = 13.0 Hz, J1b,2 = 9.5 Hz, H-1b), 4.01 (1H, t, J2,3
=
J3,4 = 1.51 Hz, H-3), 4.24–4.34 (4H, m, 4· ArCH2), 4.38
(1H, s, NH amine), 4.42 (1H, t, J3,4 = J4,5 = 1.5 Hz, H-
4), 4.44–4.48 (1H, m, H-2), 4.51 (1H, d, J4,5 = 1.5 Hz,
H-5), 7.19–7.27 (2H, m, 2· ArH), 7.41–7.73 (4H, m,
4· ArH), 8.54 (1H, t, J14, NH-amide = 6.5 Hz, NH-amide).
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmc.2008.03.045.
ESI+: [M+H]+ = 493.1 [35Cl–35Cl–35Cl–35Cl].
References and notes
1. Verlinde, C. L. M. J.; Hannaert, V.; Blonski, C.; Willson,
´ ´
5.1.14. Activity screens against Tb PFK. In a 1 mL cuv-
ette was placed an inhibitor solution (40 lL, made up
with 0.1 M triethanolamine buffer, pH 8.0), 12.5 lg/
mL PFK (16 lL), made up to 760 lL with 0.1 M trieth-
anolamine, pH 8.0 and was incubated for 2 min. The
reaction was started by the addition of the substrate
mix (40 lL, 40 mM MgCl2, 40 mM F6P, 20 mM ATP,
6 mM NADH, 18 U/mL aldolase [Sigma], 20 U/mL tri-
osephosphate isomerase [Sigma] and 34 U/mL glycerol-
3-phosphate dehydrogenase [Sigma]), gently agitated
and the decrease in absorbance at 340 nm measured
for 2 min (using Lambda Bio). The initial rate was then
calculated using UV kinlab. A control rate with no
inhibitor present was also determined. The rate for each
inhibitor assay was expressed as a percentage of the con-
trol assay. All compounds considered as hits were
screened against the coupling enzymes alone to avoid
a false positive result. All compounds were found to in-
hibit PFK only.
M.; Perie, J. J.; Fothergill-Gilmore, L. A.; Opperdoes, F.
R.; Gelb, M. H.; Hol, W. G. J.; Michels, P. A. M. Drug
Res. Updates 2001, 4, 50.
2. Albert, M. A.; Haanstra, J. R.; Hannaert, V.; Van Roy, J.;
Opperdoes, F. R.; Bakker, B. M.; Michels, P. A. J. Biol.
Chem. 2005, 280, 28306.
3. Fernandes, A. P.; Nelson, K.; Beverley, S. M. Proc. Natl.
Acad. Sci. U.S.A. 1993, 90, 11608.
4. Fothergill-Gilmore, L. A.; Michels, P. A. M. Prog.
Biophys. Mol. Biol. 1993, 59, 105.
5. Michels, P. A. M.; Chevalier, N.; Opperdoes, F. R.; Rider,
M. H.; Rigden, D. J. Eur. J. Biochem. 1997, 250, 698.
6. Van Schaftingen, E.; Opperdoes, F. R.; Hers, H. G. Eur. J.
Biochem. 1985, 153, 403.
7. Martinez-Oyanedel, J.; McNae, I. W.; Nowicki, M. W.;
Keillor, J. W.; Michels, P. A. M.; Fothergill-Gilmore, L.
A.; Walkinshaw, M. D. J. Mol. Biol. 2007, 366, 1185.
8. Claustre, S.; Denier, C.; Lakhdar-Ghazal, F.; Lougare, A.;
´ ´
Lopez, C.; Chevalier, N.; Michels, P. A. M.; Perie, J.;
Willson, M. Biochemistry 2002, 41, 10183.
9. Rigden, D. J.; Phillips, S. E. V.; Michels, P. A. M.;
Fothergill-Gilmore, L. A. J. Mol. Biol. 1999, 291, 615.
5.1.15. Activity screens against Lm PyK. To a 1 mL cuv-
ette was added the substrate mix (500 lL, 50 mM trieth-
anolamine buffer, pH 7.2, 50 mM KCl, 6 mM MgSO4,
0.6 mM ADP, 0.84 mM NADH, 27.6 units/mL lactate
dehydrogenase [Roche]), 0.04 mg/mL PyK (10 lL), an
inhibitor solution (50 lL, made up with 0.1 M trietha-
nolamine, pH 7.5) and made to 992 lL by the addition
of water. The mixture was incubated for 2 min before
the reaction was started by the addition of 50 lM PEP
(50 lL). The mixture was gently agitated and the de-
crease in absorbance at 340 nm was measured for
2 min (using Lambda Bio). The data were processed in
the same manner as for the screen against PFK. All
compounds considered as hits were screened against
´
´ ´
10. Claustre, S.; Bringaud, F.; Azema, L.; Baron, R.; Perie, J.;
Willson, M. Carbohydr. Res. 1999, 315, 339.
11. Horton, D.; Philips, K. D. Carbohydr. Res. 1973, 30, 367;
Piper, I. M.; MacLean, D. B.; Kvarnstrom, I.; Szarek, W.
A. Can. J. Chem. 1983, 61, 2721.
12. Booth, R. J.; Hodges, J. C. J. Am. Chem. Soc. 1997, 119,
4882.
13. Booth, R. J.; Hodges, J. C. Acc. Chem. Res. 1999, 32, 18.
14. Kaldor, S. W.; Siegel, M. G.; Fritz, J. E.; Dressman, B. A.;
Hahn, P. J. Tetrahedron Lett. 1996, 37, 7193.
15. Davis, N. J.; Flitsch, S. L. Tetrahedron Lett. 1993, 34, 1181.
16. Furka, A.; Sebestyen, F.; Asgedom, M.; Dibo, G. Int. J.
Pept. Prot. Res. 1991, 37, 487.
17. Han, S.-Y.; Kim, Y.-A. Tetrahedron 2004, 60, 2447.
18. Shirakihara, Y.; Evans, P. R. J. Mol. Biol. 1988, 204, 973.

M. W. Nowicki et al. / Bioorg. Med. Chem. 16 (2008) 5050–5061
5061
19. Fothergill-Gillmore, L. A.; Rigden, D. J.; Michels, P. A.;
Phillips, S. E. Biochem. Soc. Trans. 2000, 28, 86.
20. Characterisation data for this series can be found in the
supplementary data. Only characterisation for the main
compounds 2m and 3m are given here.
21. Details of mass directed purification methods can be
found in the supplementary material.
22. Characterisation data for this series can be found in
the supplementary data. Only characterisation for the
compound 7e is given here. Due to low yields and
amounts of material, characterisation for this part was
minimal.
23. Characterisation data for this series can be found in the
supplementary data. Only characterisation for the com-
pound 8e is given here.
´
24. Hoet, S.; Stevigny, C.; Block, S.; Opperdoes, F.; Colson,
P.; Baldeyrou, B.; Lansiaux, A.; Bailly, C.; Quentin-
Leclercq, J. Planta Med. 2004, 70, 407.