Enediol mimics as inhibitors of the D-arabinose 5-phosphate isomerase (KdsD) from Francisella tularensis

Article abstract of DOI:10.1016/j.bmcl.2010.12.066

We explored the d-arabinose 5-phosphate isomerase (KdsD, E.C. 5.3.1.13) from Francisella tularensis, a highly infectious Gram-negative pathogen that has raised concern as a potential bioweapon, as a target for the development of novel chemotherapeutics. F

Full text of DOI:10.1016/j.bmcl.2010.12.066

Bioorganic & Medicinal Chemistry Letters 21 (2011) 2679–2682
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Enediol mimics as inhibitors of the D-arabinose 5-phosphate isomerase
(KdsD) from Francisella tularensis
Alejandra Yep, Roderick J. Sorenson, Michael R. Wilson, H. D. Hollis Showalter, Scott D. Larsen,
⇑
Paul R. Keller , Ronald W. Woodard
Department of Medicinal Chemistry, University of Michigan, 428 Church Street, Ann Arbor, MI 48109-1065, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
We explored the D-arabinose 5-phosphate isomerase (KdsD, E.C. 5.3.1.13) from Francisella tularensis, a
Received 2 November 2010
Revised 10 December 2010
Accepted 15 December 2010
Available online 19 December 2010
highly infectious Gram-negative pathogen that has raised concern as a potential bioweapon, as a target
for the development of novel chemotherapeutics. F. tularensis KdsD was expressed in Escherichia coli from
a synthetic gene, purified, and characterized. A group of hydroxamates designed to be mimics of the puta-
tive enediol intermediate in the enzyme’s catalytic mechanism were prepared and tested as inhibitors of
F. tularensis KdsD. The best inhibitor, which has an IC50 of 7
reported to date.
l
M, is the most potent KdsD inhibitor
Keywords:
D-Arabinose 5-phosphate isomerase (A5PI)
Ó 2011 Published by Elsevier Ltd.
KdsD (EC 5.3.1.13)
Francisella tularensis
Enediol
Inhibitor design
Francisella tularensis, a Gram-negative bacterium from the gam-
ma subdivision of the proteobacteria, is a highly infectious patho-
gen that has raised concern as a potential bio-weapon.1 While
standard antibiotics like streptomycin, tetracyclines and fluoro-
quinolones are effective against natural strains of F. tularensis,
the potential for resistance, natural or artificial, creates a need
for an effective antibiotic that works through a novel target.
One source of novel antibiotic targets is the biosynthetic path-
way of the outer membrane component lipopolysaccharide (LPS).
LPS biosynthesis is an attractive target because it is both unique
to Gram-negative bacteria and essential to their viability.2 Defects
in LPS biosynthesis also increase the susceptibility of Gram-nega-
tive bacteria to standard antibacterials because they increase the
permeability of the outer membrane to these drugs. A key compo-
5-phosphate (A5P), which is catalyzed by the enzyme D-arabinose
5-phosphate isomerase (KdsD, E.C. 5.3.1.13). Cells of the model
Gram-negative bacterium Escherichia coli lacking this enzyme re-
quire exogenous A5P for LPS synthesis and cellular growth.3 The
lack of A5P in mammalian serum makes KdsD essential for viability
in the host, and a good target for the discovery of novel antibiotics
targeting F. tularensis.
We took a classic approach to finding inhibitors of KdsD from F.
tularensis by testing compounds that mimic the putative high-en-
ergy 1,2-enediol intermediate in the enzymatic inter-conversion of
Ru5P and A5P (Fig. 1). This approach is not without precedent. Bing-
ham et al.4 used aldonic acids and alditols as mimics of the enediol
intermediate when searching for inhibitors of E. coli KdsD. Their
most potent inhibitor was a four-carbon aldonic acid (1, Fig. 2), for-
nent of LPS is the unique eight-carbon sugar 3-deoxy-
D-manno-oct-
mally derived from D-erythrose 4-phosphate, which had an IC50 of
ulosonic acid 8-phosphate (KDO). KDO serves as the bridge
between the membrane anchor of LPS, known as lipid A, and the
inner oligosaccharide core of LPS. Biosynthesis of KDO requires
four steps starting from the pentose pathway intermediate
D
-ribulose 5-phosphate (Ru5P). The first, committed step in the
biosynthesis of KDO is the isomerization of Ru5P to D-arabinose
⇑
Corresponding author. Tel.: +1 734 763 7366; fax: +1 734 647 8430.
E-mail addresses: ayep@umich.edu (A. Yep), sorensr@umich.edu (R.J. Sorenson),
mwwils@umich.edu (M.R. Wilson), showalh@umich.edu (H.D. HollisShowalter),
sdlarsen@umich.edu (S.D. Larsen), prkeller@umich.edu (P.R. Keller), rww@umich.
edu (R.W. Woodard).
Present address: Department of Radiology, University of Michigan Medical
School, BSRB A688, Ann Arbor, MI 48109-2200, USA.
Figure 1. Isomerization catalyzed by KdsD and the enzyme-bound enediol
intermediate.
0960-894X/$ - see front matter Ó 2011 Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2010.12.066

2680
A. Yep et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2679–2682
Figure 2. Chemical structures referred to in this study.
220
l
M. This is the most potent inhibitor of KdsD described in
In addition to these classic cases, hydroxamates have been used
as chemical probes with a number of other sugar isomerases, in
some cases proving to be potent inhibitors of enzymes derived
from pathogens. Pastiet al.9 and Dardonville et al.10 synthesized
and tested hydroxamates as inhibitors of Trypanosoma brucei 6-
phosphogluconate dehydrogenase (6PGDH), which catalyzes the
oxidation and subsequent decarboxylation of 6-phosphogluconate
to form Ru5P. Pasti et al.9 found that 5-carbon hydroxamate 2
the literature. To improve upon this, we chose to explore hydroxa-
mic acids, which more closely resemble the putative enediol
intermediate.
Hydroxamates have been used as enediol mimics in many similar
systems. In three classic cases, hydroxamates were used to help elu-
cidate the structure of the active site and the enzyme mechanism.
Collins5 first demonstrated that phosphoglycolohydroxamate, a
two-carbon hydroxamate, inhibits triose phosphate isomerase
(Fig. 2) inhibited T. brucei 6PGDH with an IC50 of 5.8 lM. Dardon-
(TIM) with a Ki of 4
lM. Details of the active site and the mechanism
ville et al.10 found that 4-carbon hydroxamates 3–5 (Fig. 2) pro-
vided Ki values below 100 nM, and amide 6 (Fig. 2) a Ki of
of catalysis were elucidated with the help of X-ray crystal structures
(e.g., 2VXN6) of TIM–phosphoglycolohydroxamate complexes. The
inter-conversion of dihydroxyacetone phosphate and glyceralde-
hyde 3-phosphate phosphate catalyzed by TIM is thought to proceed
via a proton transfer mechanism, through a cis-enediol intermedi-
1.52 lM versus T. brucei 6PGDH. Hydroxamate inhibitors have also
been found to inhibit RpiB from Mycobacterium tuberculosis.11,12
RpiB catalyzes the inter-conversion of ribose 5-phosphate (R5P)
and Ru5P, but its putative physiological role is to inter-convert
the six-carbon sugars allose 5-phosphate and allulose 5-phosphate.
ate. Allen et al.7 found
D-threonohydroxamic acid, a four-carbon
hydroxamate, to be a slow-binding inhibitor of xylose isomerase,
with a Ki of 100 nM. As with TIM, the details of the xylose isomerase
active site were elucidated with the help of an X-ray crystal struc-
ture (2GYI)7 of the hydroxamate-xylose isomerase complex. Xylose
isomerase catalyzes the inter-conversion of five-carbon sugars xy-
lose and xylulose, but is also used industrially to inter-convert the
six-carbon sugars glucose and fructose. This metallo-enzyme fol-
lows a slightly different mechanistic path than TIM, involving the
The four-carbon hydroxamate 4 (Fig. 2) had a Ki of 57 lM against
the M. tuberculosis RpiB, while the aldonic acid 1 had a Ki of
1.7 mM.11,12 A five-carbon hydroxamate mimic of the putative ene-
diol intermediate in the inter-conversion of allose to allulose was a
poorer inhibitor (Ki of 400 lM) than 4, but still better than aldonic
acid 1. These precedents reveal that hydroxamates can be potent
inhibitors of carbohydrate isomerases, and that the best starting
point is to prepare hydroxamates one carbon shorter than the
shortest sugar substrate. In order to find a starting point for F.
tularensis KdsD inhibitor design, we decided to prepare and test
the four most potent inhibitors of T. brucei 6PGDH, the four-carbon
analogs 3–6 (Fig. 2).
transfer of hydride. Arsineiva et al.8 used 5-phospho-
D-arabinohydr-
oxamic acid (2, Fig. 2) to study the mechanism of phophoglucose
isomerase (PGI), a cytosolic enzyme that catalyzes the reversible
isomerization of glucose 6-phosphate to fructose 6-phophate. The
Ki for 5-phospho-
D
-arabinohydroxamic acid versus rabbit PGI is
Compounds 3–6 were synthesized according to procedures out-
lined by Dardonvilleet al.10 The reaction scheme is presented in
Figure 3; the full experimental details relating to the synthesis of
these compounds are delineated in the Supplementary data.
100 nM.8 An X-ray crystal structure of the PGI–hydroxamate
complex (1KOJ)8 suggests a proton transfer mechanism with a cis-
enediol intermediate similar to that of TIM.
Figure 3. Synthesis of hydroxamate analogs 3–6.

A. Yep et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2679–2682
2681
Table 1
Biological activity of compounds 3–6
Compound
F. tularensis KdsD IC50
(lM)
T. brucei 6PGDH Kia
(lM)
M. tuberculosis RpiB Kib
(lM)
E. coli MIC (mM)
3
4
5
6
3000 900
0.035
0.01
0.08
1.52
—
—
52
18
35
7
6
1
2
57
10
>30,000
a
From Dardonville et al.10
From Roos et al.12
b
Difficulties in reproducibly generating a key protected phosphate
intermediate in the synthesis of inhibitor 6 required development
of a novel work-around. Rather than use the tribenzyl phosphite
and iodine procedure employed by Dardonville et al.10 to introduce
the tribenzylated phosphate in compound 10, we employed dib-
enzyldiisopropylphosphoramidite and tetrazole to form intermedi-
ate 9 (Fig. 3), which was oxidized without isolation using tert-butyl
hydroperoxide. The yield of compound 10 (78% from compound 8)
was slightly superior to that reported by Dardonville et al. (73%),
but had the additional advantage of being more reproducible.
The F. tularensis KdsD protein used in this study was derived
from a commercially prepared synthetic gene based upon the pub-
lished F. tularensis genome sequence. This synthetic gene was sub-
cloned into the expression vector pT7LOH,13 creating a plasmid
that encodes KdsD with an N-terminal hexahistidine tag. This plas-
mid was used to transform E. coli BL21(DE3) cells that were then
used to produce multi-milligram quantities of metal-affinity puri-
fied KdsD using standard growth, expression, lysis and purification
techniques (see Supplementary data). The resulting protein was
substantially pure, judged by SDS–PAGE.
with the protein. The presence of the acetyl groups led to an eight-
fold decline in potency against T. brucei 6PGDH,10 but a fivefold
increase in potency toward F. tularensis KdsD. The amide 6, while
it maintains the substrate-like positioning of the hydroxyl groups,
lacks the enediol mimicry of the hydroxamate. This led to easy-to-
rationalize losses in potency against T. brucei 6PGDH (150-fold),10
and M. tuberculosis RpiB (>750-fold),12 but a 3.5-fold increase in
potency against F. tularensis KdsD.
To further investigate the properties of these inhibitors, we
tested the ability of the two most potent compounds, 5 and 6, to
inhibit the growth of E. coli in culture.14 We used this model
Gram-negative organism for the sake of convenience, as F. tularen-
sisis not generally available for these types of studies. As shown in
Table 1, the minimum inhibitory concentrations (MICs) were in the
tens of millimolar range, approximately 1000-fold their IC50s for F.
tularensis KdsD.
Compound 5 now represents the most potent KdsD inhibitor re-
ported to date, which is clearly a step in the right direction. The
limited selectivity and cytotoxicity data available, however, sug-
gest that work will be required to convert compound 5 into a solid
lead. Its Ki against T. brucei 6PGDH is 80 nM and its Ki against sheep
6PGDH is 360 nM.10 The in vitro antiparasitic activity and the cyto-
The purified F. tularensis KdsD was assayed for isomerase activ-
ity in the reverse of the physiological direction, using A5P as sub-
strate and determining the product Ru5P using the discontinuous
cysteine-carbazole colorimetric method.3 F. tularensis KdsD had
toxicity against L6 cells were 229 l
M or greater—at least 103 high-
er than the enzyme inhibitory level10—but perhaps lower than the
52 mM E. coli MIC (Table 1). Clearly, a round or two of optimization
is necessary before one would expect to find selectivity. Informa-
tion that would be helpful for optimizing the inhibitors described
here is currently limited.
an A5P Km of 380 80 l
M and a kcat of 141 9 sÀ1 at 37 °C using
this assay. These values are very similar to values (Km = 610
lM
and kcat = 157 sÀ1) published for E. coli KdsD.3
Compounds 3–6 were then assayed for inhibition of KdsD at
37 °C using the same discontinuous cysteine-carbazole colorimet-
ric method, adapted to a 96-well microplate, used to characterize
the enzyme. The full details are presented in the Supplementary
data. Briefly, compound and KdsD (100 nM final concentration)
were mixed in buffer at pH 8.5, and then warmed to 37 °C. An
equal volume of A5P (1 mM final concentration), in the same buf-
fer at 37 °C, was added to initiate the reaction. After a 5 min reac-
tion time, the assay was stopped with sulfuric acid, transferred to
a separate plate, mixed with the sulfuric acid–cysteine–carbazole
reagent and allowed to stand at ambient temperature for 3 h for
color development. Absorbance at 540 nm was then measured.
IC50s for each of the compounds studied are presented in Table 1.
The four-carbon hydroxamate 4 inhibits F. tularensis KdsD with
Unlike T. brucei 6PGDH11 and M. tuberculosis RpiB,12 there are
no crystal structures of F. tularensis KdsD. The closest model avail-
able is the very recently reported X-ray crystal structure of a frag-
ment, the sugar isomerase domain, of an inactive site-directed
mutant of E. coli KdsD.15 While sufficient for making predictions
about which residues are involved in catalysis, this model is insuf-
ficient to explain the inhibition results presented here. A high-res-
olution structure of the full-length, wild-type F. tularensis KdsD
protein containing a small molecule in the active site is needed
to make rapid progress. In the absence of this structure, however,
there are still steps that can be taken. Although the hydroxamates
in this study did not inhibit KdsD as potently as they inhibit T.
brucei 6PGDH, this does not rule out the use of other enediol mim-
ics to increase potency. Other enediol mimics include the oxime
used by Bearne and Blouin16 as an inhibitor of glucosamine-6-
phosphate synthase, and the reverse-hydroximate found in the
antibacterial fosmidomycin,17 an inhibitor of bacterial 1-deoxy-
a potency (IC50 35
M. tuberculosis RpiB (Ki = 57
6PGDH (Ki = 0.01
M).10 The acetonide 3, in which the sugar hy-
l
M), which is more in line with its inhibition of
l
M)12 than its inhibition of T. brucei
l
droxyl groups are masked in a five-membered ring, is 35-fold less
potent than 4 against F. tularensis KdsD, while its potency de-
creases only 3.5-fold toward T. brucei 6GPDH.10 The potency ratio
we observed toward KdsD seems more reasonable than the ratio
observed toward T. brucei 6GPDH, and Dardonville et al.10 appar-
ently agree, since they attributed the small ratio they observed
to hydrolysis of the acetonide. The diacetate 5 and the amide 6
were both slightly more potent than the hydroxamate 4, a result
that was unexpected. The hydroxyl groups on hydroxamate 4 are
positioned analogous to those in the substrate Ru5P, and the acetyl
groups in 5 would be expected to disrupt the natural interactions
D-xylulose 5-phosphoate reductoisomerase. Finally, the assay
described here could be used to screen individually synthesized
analogs of 5, or collections of randomly selected compounds as
a means of identifying promising chemical matter for drug discov-
ery efforts.
Acknowledgements
This work was supported by National Institute of Health Grant
AI-061531 (to R.W.W.). The authors would like to thank Ms. Yafei

2682
A. Yep et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2679–2682
5. Collins, K. D. J. Biol. Chem. 1974, 249, 136.
Jin for performing the synthesis of compound 5 and Dr. Tod P. Hol-
ler for his helpful suggestions and for editing the manuscript.
6. Alahuhta, M.; Wierenga, R. K. Proteins 2010, 78, 1878.
7. Allen, K. N.; Lavie, A.; Petsko, G. A.; Ringe, D. Biochemistry 1995, 34, 3742.
8. Arsenieva, D.; Hardré, R.; Salmon, L.; Jeffrey, C. J. Proc. Natl. Acad. Sci. U.S.A.
2002, 99, 5872.
9. Pasti, C.; Rinaldi, E.; Cervellati, C.; Dallocchio, F.; Hardré, R.; Salmon, L.; Hanau,
S. Bioorg. Med. Chem. 2003, 11, 1207.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.12.066.
10. Dardonville, C.; Rinaldi, E.; Barrett, M. P.; Brun, R.; Gilbert, I. H.; Hanau, S. J.
Med. Chem. 2004, 47, 3427.
11. Burgos, E.; Roos, A. K.; Mowbray, S. L.; Salmon, L. Tetrahedron Lett. 2005, 46,
3692.
12. Roos, A. K.; Burgos, E.; Ericsson, D. J.; Salmon, L.; Mowbray, S. L. J. Biol. Chem.
2005, 280, 6416.
References and notes
13. Muda, M.; Worby, C. A.; Simonson-Leff, N.; Clemens, J. C.; Dixon, J. E. Biochem. J.
2002, 366, 73.
14. Wiegand, I.; Hilpert, K.; Hancock, R. E. Nat. Protocol 2008, 3, 163.
15. Gourlay, L. J.; Sommaruga, S.; Nardini, M.; Sperandeo, P.; Dehò, G.; Polissi, A.;
Bolognesi, M. Protein Sci. 2010.
16. Bearne, S. L.; Blouin, C. J. Biol. Chem. 2000, 275, 135.
17. Jomaa, H.; Wiesner, J.; Sanderbrand, S.; Altincicek, B.; Weidemeyer, C.; Hintz,
M.; Türbachova, I.; Eberl, M.; Zeidler, J.; Lichtenthaler, H. K.; Dominique Soldati,
D.; Beck, E. Science 1999, 285, 1573.
1. Dennis, D. T.; Inglesby, T. V.; Henderson, D. A.; Bartlett, J. G.; Ascher, M. S.;
Eitzen, M.; Fine, A. D.; Friedlander, A. M.; Hauser, J.; Layton, M.; Lillibridge, S. R.;
McDade, J. E.; Osterholm, M. T.; O’Toole, T.; Parker, G.; Perl, T. M.; Russell, P. K.;
Tonat, K. JAMA 2001, 285, 2763.
2. Raetz, C. R. H.; Whitfield, C. Ann. Rev. Biochem. 2002, 71, 635.
3. Meredith, T. C.; Aggarwal, P.; Mamat, U.; Lindner, B.; Woodard, R. W. ACS Chem.
Biol. 2006, 1, 33.
4. Bingham, E. C.; Gragg, C. E.; Hall, W. R.; Kelsey, J. E.; Mallory, W. R.; Richardson,
D. C.; Benedict, C.; Ray, P. H. J. Med. Chem. 1984, 27, 717.