Design and structure-activity relationships of potent and selective inhibitors of undecaprenyl pyrophosphate synthase (UPPS): Tetramic, tetronic acids and dihydropyridin-2-ones

Article abstract of DOI:10.1016/j.bmcl.2008.02.009

Based on a pharmacophore hypothesis substituted tetramic and tetronic acid 3-carboxamides as well as dihydropyridin-2-one-3-carboxamides were investigated as inhibitors of undecaprenyl pyrophosphate synthase (UPPS) for use as novel antimicrobial agents. S

Full text of DOI:10.1016/j.bmcl.2008.02.009

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1840–1844
Design and structure–activity relationships of potent
and selective inhibitors of undecaprenyl pyrophosphate
synthase (UPPS): Tetramic, tetronic acids
and dihydropyridin-2-ones
Stefan Peukert,* Yingchuan Sun, Rui Zhang, Brian Hurley, Mike Sabio, Xiaoyu Shen,
Christen Gray, JoAnn Dzink-Fox, Jianshi Tao, Regina Cebula and Sompong Wattanasin
Novartis Institutes for BioMedical Research, Inc., 250 Massachusetts Avenue, Cambridge, MA 02139, USA
Received 17 January 2008; accepted 7 February 2008
Available online 10 February 2008
Abstract—Based on a pharmacophore hypothesis substituted tetramic and tetronic acid 3-carboxamides as well as dihydropyridin-2-
one-3-carboxamides were investigated as inhibitors of undecaprenyl pyrophosphate synthase (UPPS) for use as novel antimicrobial
agents. Synthesis and structure–activity relationship patterns for this class of compounds are discussed. Selectivity data and anti-
bacterial activities for selected compounds are provided.
Ó 2008 Elsevier Ltd. All rights reserved.
Undecaprenyl pyrophosphate synthase (UPPS) is an en-
zyme essential for bacterial viability. It catalyzes cis-
double bond formation during sequential condensation
of eight isopentenyl pyrophosphate (IPP) molecules with
farnesyl pyrophosphate (FPP) to form C55 undecaprenyl
pyrophosphate, which is the lipid carrier for the precur-
sors of various cell wall structures, such as peptidogly-
can, teichoic acids, and O-antigens.1 This critical
biological function as well as an active site amenable
to small molecule inhibition, as revealed by crystallo-
graphic structures, make UPPS an attractive target for
the discovery of novel antibacterial agents.2
In the present study, we describe the first potent and selec-
tive inhibitors of UPPS.3 The work profited from the re-
Figure 1. Structure of undecaprenyl pyrophosphate synthase (UPPS)
cently published co-crystal structure of UPPS with the
natural substrate farnesyl pyrophosphate in the active site
from E. coli in complex with farnesyl pyrophosphate (in pink color; H-
bonds in orange). Tetramic-acid derivative 1a (yellow color) is docked
of the enzyme.2b Based on the crystal structure, the pyro-
into the binding site occupied by FPP, showing two H-bonds to Arg39
and an internal H-bond (in blue).
phosphate head of FPP is bound to the backbone NHs of
Gly29 and Arg30 as well as the side chains of Asn28,
Arg30, and Arg39 through hydrogen bonding (Fig. 1).
The hydrocarbon moiety of FPP interacts with the side
chains of hydrophobic amino acids. The micromolar
inhibitor of UPPS, tetramic-acid derivative 1a, was iden-
tified in a high throughput screening and can be docked
into the active site occupying the FPP binding site (Fig. 1).
Keywords: UPPS inhibitor; Undecaprenyl pyrophosphate synthase;
Antibacterial.
The analysis of this binding mode resulted in the
hypothesis that two hydrogen acceptors and a hydro-
*
Corresponding author. Tel.: +1 617 871 3644; fax: +1 617 871
4081; e-mail: stefanpeukert@yahoo.com
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.02.009

S. Peukert et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1840–1844
1841
Enantioselective synthesis of 3-acyl tetramic acids 1
starting from enantiomerically pure amino acids was
not pursued because of reported partial racemization
under the cyclization conditions.6 Instead, racemates
could be separated into the pure enantiomers
(ee > 98%) by chiral chromatography on a Chiralpak
AD column. Sodium salts of 1 were isolated by precipi-
tation after the addition of 0.95 equiv aqueous 1 M
NaOH to a solution of the acid in methanol.
a
b
OH
O
O
N
N
H
O
H
1a
OH
OH
OH
O
O
O
R1
R1
R2
R2
R2
N
N
N
O
N
N
H
H
H
O
O
O
R1
H
2
1
3
The synthesis of 3-acyl tetronic acids 3 employed the
methodology by Igglessi-Markopoulou.7 a-Hydroxy
acids 9 were O-acylated with acetic anhydride followed
by C-acylation between dimethyl malonate and the N-
hydroxybenzotriazole ester of the O-protected a-hydro-
xy acid 10. Cyclization to compounds 12 was achieved
by treatment with 20% HCl in dioxane for 16 h. Amide
bond formation with amines in toluene under reflux
OH O
R2
OH O
R1
R2
N
H
N
H
R1
N
H
O
N
O
5
H
4
Figure 2. (a) Chemical structure of 1a with the 4-hydroxy and amide
bond oxygen atoms as hydrogen acceptors and the 4-hexyloxyphenyl
group as hydrophobic contact. (b) Classes of potent UPPS inhibitors
identified using the pharmacophore hypothesis.
CO2H
O
CO2H
R1
10
a
b
R1
HO
O
phobic group serve as pharmacophoric points (Fig. 2a)
whereas the ring carbonyl oxygen together with the exo-
cyclic amide NH forms a hydrogen bond to stabilize the
conformation of the molecule.
9
CO2Me
OH
OH
O
O
CO Me
2 c
R1
R1
d
R2
CO2Me
R1
N
O
O
O
H
O
O
This hypothesis proved fruitful in our endeavour to de-
sign and synthesize novel UPPS inhibitors. Along with
tetramic acids 1, N-alkylated tetramic acid 2, tetronic
acids 3 and the two regioisomeric dihydropyridin-2-ones
4 and 5 match this pharmacophore and demonstrate po-
tent inhibition of UPPS (Fig. 2b).
O
3a-d
11
12
Scheme 2. Reagents and conditions: (a) acetic anhydride, pyridine,
16 h, 100%; (b) 1—HOBt, DCC, THF, 0 °C, 16 h; 2—NaH, dimethyl
malonate, 0 °C to rt, 3 h; (c) dioxane, 20% HCl, 16 h, 60% (b + c); (d)
R2-NH2, toluene, reflux, 1 h, 40–60%.
Synthesis: Tetramic acids (1 and 2) were prepared in
three steps from commercially available racemic a-ami-
no acid methyl esters 6. Acylation with methyl malonyl
chloride provided the intermediates 7 which were sub-
jected to Lacey–Dieckmann cyclization conditions to
yield 3-methoxycarbonyl tetramic acids 8 as building
blocks in good yields.4 Formation of 3-carboxamides
was best achieved with microwave heating at tempera-
tures between 100 and 120 °C in THF or ethanol as
solvent. Higher temperatures caused significant decar-
boxylation and formation of vinylogous amides.5 In
the case of 1g (R1 = 2-imidazolylmethyl) removal of
the trityl-protecting group with trifluoroacetic acid com-
pleted the synthesis (Scheme 1).
CO2Me
CO2Me
R1
O
O
a
b
R1
13
OH O
H2N
R1
MeO
N
H
14
R2
O HN
OH O
R2
c
OMe
O
OH
N
H
N
H
O
R1
R1
N
H
N
H
O
15
4a - m
4'a - m
Scheme 3. Reagents and conditions: (a) methyl malonyl chloride,
CH2Cl2, Et3N, rt, 2 h, 76–100%; (b) NaOMe, MeOH, reflux, 2 h, 12–
77%; (c) R2-NH2, THF, microwave synthesizer, 120–150 °C, 5–8 min,
40–80%.
O
O
CO2Me
R1
a
b
CO2Me
R1
a
MeO
c
N
R
HN
R
NC CO2Me
NC CO2Me
6
7
16
17
OH
OH
O
O
OH O
R1
R1
R2
R1
R1
R2
b
CO2Me
see Scheme 3
N
OMe
N
H
N
N
H
R
O
H2N
O
R
N
H
O
5a, b
8
1a-l, 2
18
Scheme 1. Reagents and conditions: (a) methyl malonyl chloride,
CH2Cl2, Et3N, rt, 2 h, 72–98%; (b) NaOMe, MeOH, reflux, 2 h, 76–
99%; (c) R2-NH2, THF or EtOH, microwave synthesizer, 100–120 °C,
5–8 min, 40–80%, for 1g: then TFA, rt, 2 h.
Scheme 4. Reagents and conditions: (a) piperidine, toluene, benz-
aldehyde, reflux, 4 h, 74%; (b) for R1 = PhCH2: Pearlman’s cat.,
EtOH, concd HCl, 50 psi H2, rt, 3 d, 91%; for R1 = cyclohexyl-CH2:
PtO2, EtOH, concd HCl, 50 psi H2, rt, 3 d, 77%.

Table 1. In vitro activity against S. pneumoniae (sp) UPPS, human (hu) FPPS (IC50s, lM) and minimal inhibitory concentration (MIC, lg/mL)9
X
R1
R2
spUPPS
huFPPS
Enterococcus faecalis
Staphylococcus aureus
S. pneumoniae
OH
O
R1
R2
N
X
H
O
1a
1b
1c
1d
1e
1f
NH
NH
NH
NH
NH
NH
NH
NH
NH
PhCH2
PhCH2
4-Hexyloxyphenyl
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Biphenyl
19
1.8
58
>100
8
0.5
1
H
Ph
3.5
0.16
1.3
4.2
>50
0.12
>100
>100
>50
4
2
2
1
1
8
PhCH2CH2
Isobutyl
4
4
2
16
0.5
16
1g
1h
1i
2-Imidazolyl-methyl
Methoxymethyl
PhCH2CH2
>100
>50
16
8
4
4
2
4
1j
NH
PhCH2CH2
0.80
N
1k
1l
NH
NH
PhCH2CH2
PhCH2CH2
4-Fluorophenyl
Cyclohexyl
2.5
35
64
32
16
32
16
64
OH
O
2
0.3
8
1
0.5
N
N
H
Ph
O
3a
3b
3c
3d
O
O
O
O
Ph
Ph
4-Chlorophenyl
3,5-Dichlorophenyl
4-Cyclohexylphenyl
4-Biphenyl
24
>50
>50
>50
>64
32
>64
32
64
4
2.0
0.5
1.1
PhCH2
PhCH2
32
32
64
>128
1
4

S. Peukert et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1840–1844
1843
conditions furnished the tetronic-acid derivates
(Scheme 2).
3
zyl and cyclohexylmethyl) were easily prepared from the
Knoevenagel product 17 by hydrogenation with either
Pearlman’s catalyst or PtO2 under otherwise identical
conditions.
The less common carboxamides of dihydropyridin-2-
one-3-carboxamides 4 were synthesized in a similar fash-
ion to the tetramic acids 1 starting from b-amino acid es-
ters 13 (Scheme 3). Reaction with methyl malonyl
chloride provided the malonyl amides 14 which were
condensed to the esters 15.8 The amides 4 were formed
by reaction with amines in tetrahydrofuran in varying
yields under microwave conditions at higher tempera-
tures (120–150 °C) compared to the tetramic acids 1.
Enantiomerically pure compounds 4 were obtained
starting from commercially available (S)- or (R)-b-ami-
no acid esters 13.
Results and discussion: Table 1 shows the structure–
activity relationship (SAR) for a number of substituents
R1 and R2 within the tetramic (X@NH) and tetronic
acid series (X@O). The potency of the initial hit 1a could
be improved by changing the 4-hexyloxyphenyl substitu-
ent to the conformationally more restricted 4-cyclo-
hexylphenyl group (1b). Hydrophobic groups at the R1
position proved beneficial (1d–f), whereas the unsubsti-
tuted compound 1c showed significantly reduced en-
zyme inhibition. Polar R1 substituents (1g, h) are less
tolerated and resulted in higher IC50 values. Interest-
ingly, the enantiomers of the active compound 1e had al-
most identical enzyme activity (data not shown).
Compounds 4 exist in DMSO-d6 as a ꢀ1:1 mixture of
the two tautomers 4 and 40 with different orientation
of the amide substituent R2.8 In addition, these com-
pounds are weak acids with a pKa of ꢀ6.5–7 compared
to the tetramic and tetronic acids 1, 2 and 3 with pKa’s
of 3–3.5.
Sodium salts of compounds 1 provided similar potency
as the free acids. Optimization of R2 resulted in aryl-
phenyl groups (e.g., 1i) or heterocycloalkyl-phenyl
groups (e.g., 1j). Activity decreased with smaller aro-
matic substituents (1k) and dropped further with ali-
phatic groups (1l) at this position. The substituent R1
could be moved to the ring nitrogen (2) without loss
Regioisomeric compounds 5 were synthesized in the
same way from the corresponding b-amino acid esters
18 (Scheme 4). The two amino acid esters 18 (R1 = ben-
Table 2. In vitro activity against S. pneumoniae (sp) UPPS, human (hu) FPPS (IC50s, lM) and minimal inhibitory concentration (MIC, lg/mL)9
R1
R2
spUPPSs
huFPPS
E. faecalis
S. aureus
S. pneumoniae
OH
O
5
6
R2
R1
N
H
N
H
O
4a
4b
4c
4d
4e
4f
6-Ph
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Cyclohexylphenyl
4-Phenoxyphenyl
4-Anilinophenyl
0.2
>50
>100
>50
>50
8
4
4
(S)-6-Ph
(R)-6-Ph
6-PhCH2
6-Cyclohexyl
6-Isobutyl
6-PhCH2
6-PhCH2
0.07
0.18
0.04
0.06
0.2
64
64
64
32
64
64
4
4
4
4
32
>32
>64
16
8
>32
4
4g
4h
0.11
1.1
>50
32
O
S
4i
4j
4k
4l
6-PhCH2
8.5
16
16
32
16
64
4
32
64
32
16
64
64
16
16
64
32
8
O
6-PhCH2
0.11
0.63
0.14
0.07
8.8
>50
N
N
N
6-PhCH2
CF3
N
6-PhCH2
>50
>50
4m
5a
5b
6-PhCH2
Cyclohexyl
64
32
16
N
N
5-PhCH2
N
N
5-Cyclohexylmethyl
2.1
4

1844
S. Peukert et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1840–1844
of activity demonstrating that the ring NH is probably
not a key element for interaction with the binding site.
Tetronic acids (3a–d) showed similar enzyme activities
and the 4-cyclohexylphenyl substituent proved to be
one of the best R2 groups here, too.
References and notes
1. (a) Robyt, J. In Essentials of Carbohydrate Chemistry;
Springer-Verlag: New York, 1998; p 305; (b) Ogura, K.;
Koyama, T. Chem. Rev. 1998, 98, 1263.
2. (a) Chang, S.-Y.; Ko, T.-P.; Liang, P.-H.; Wang, A. H.-J.
J. Biol. Chem. 2003, 27, 29298; (b) Chang, S.-Y.; Ko, T.-P.;
Chen, A. P.-C.; Wang, A. H.-J.; Liang, P.-H. Protein Sci.
2004, 13, 971; (c) Guo, R.-T.; Ko, T.-P.; Chen, A. P.-C.;
Kuo, C.-J.; Wang, A. H.-J.; Liang, P.-H. J. Biol. Chem.
2005, 280, 20762.
The SAR of the dihydropyridin-2-one carboxamides 4
and 5 is summarized in Table 2. The SAR shows similar-
ity with the tetramic acids 1 but the IC50 values are typ-
ically lower. At the R1 position, hydrophobic groups are
preferred with the benzyl and cyclohexyl substituent
being the most potent groups (4a, d–f). As observed be-
fore, the difference in activity for the two enantiomers
(4b, c) is very small with a slight preference for the S-
enantiomer 4b.
3. Several UPPS inhibitors were identified by
a high
throughput screening but no structures were disclosed
(a) Li, H.; Huang, J.; Jiang, X.; Seefeld, M.; McQueney,
M.; Macarron, R. J. Biomol. Screen. 2003, 8, 712; Less
selective bisphosphonates as UPPS inhibitors are
described in (b) Guo, R.-T.; Cao, R.; Liang, P.-H.; Ko,
T.-P.; Chang, T.-H.; Hudock, M. P.; Jeng, W.-Y.; Chen,
C. K.-M.; Zhang, Y.; Song, Y.; Kuo, C.-J.; Yin, F.;
Oldfield, E.; Wang, A. H.-J. PNAS 2007, 104, 10022.
4. Lacey, R. N. J. Chem. Soc. 1954, 850.
5. Reaction of 5-benzyl-3-methoxycarbonyl tetramic acid
with aniline at 200 °C for 10 min provided 5-benzyl-4-
phenylamino-1,5-dihydro-pyrrol-2-one in 34% yield next
to the desired product 1 (40% yield):
The 4-cyclohexylphenyl group in R2 position was again
one of the best substituents (4d) but a number of other
substituents, as demonstrated with compounds 4g, j, l,
including the aliphatic cyclohexyl group (4m), showed
excellent inhibitory activity. 4-N-Imidazolyl-phenyl as a
polar residue was tolerated with a slight loss in activity
(4k). A comparison between the three compounds 4g, h
and i shows that the conformation of the R2 substituents
impacts activity: the compound with the strongly non lin-
ear 4-benzenesulfonylphenyl substituent (4i) shows less
activity than the compounds with the more linear 4-anili-
no- or 4-phenoxy-phenyl group (4g, h). Shifting the R1
substituent from the 6-position to the 5-position is toler-
ated (5a, b) but a direct comparison of compound 5a ver-
sus 4k indicates a ꢀ10-fold loss in activity. All compounds
tested showed good selectivity towards human farnesyl
pyrophosphate synthase (FPPS), a trans-prenyltransfer-
ase which is the target of bisphosphonate drugs.10
HN
Ph
OH
Ph
O
PhNH2
OMe
1
+
N
H
N
O
O
H
.
6. Poncet, J.; Jouin, P.; Castro, B.; Nicolas, L.; Boutar, M.;
Gaudemer, A. J. Chem. Soc., Perkin Trans. 1 1990, 611.
7. Athanasellis, G.; Igglessi-Markopoulou, O.; Marko-pou-
los, J. Synlett 2002, 10, 1736.
8. Weiss, R.; Presser, A.; Seebacher, W. Monatsh. Chem.
2003, 134, 1129.
9. spUPPS enzyme biochemical assay: S. pneumoniae UPPS
was expressed and purified as an N-terminal His6 fusion.
The purified enzyme was then mixed with liposome made
from Escherichia coli total lipid extract (Avanti Polar
Lipids, Inc., Alabaster, AL). The enzyme reaction was
carried out in 100 mM Tris–HCl, pH 7.3, 50 mM KCl,
1 mM MgCl2, 0.01% Triton X-100, 20 lg/mL BSA, 3 lM
FPP, 16 lM IPP, and 2 U/mL E. coli inorganic phospha-
tase. The inorganic phosphate generated in the reaction
was then quantified with Biomol Green reagent (Biomol
International, Plymouth Meeting, PA). Human FPPS
reaction was carried out under the same condition as the
one for UPPS using 8 lM DMAPP and 16 lM IPP.
MICs: The quantitative in vitro determination of mini-
mum inhibitory concentration (MICs) of novel com-
pounds against Gram-positive aerobic bacteria was
performed by microdilution broth method of testing as
recommended by the standards approved by the Clinical
and Laboratory Standards Institute (formerly, NCCLS),
Methods for Dilution Antibacterial Susceptibility Tests for
Bacteria that Grow Aerobically—7th edition. Approved
Standard Document M7-A7, Vol. 20. No. 2 CLSI, Wayne,
PA.
Most tetramic- and tetronic acids 1–3 and dihydropyrid-
iones 4–5 exhibited Gram-positive antibacterial activity
in the range of 0.5–64 lg/mL (Tables 1 and 2). A com-
parison of the three compound classes shows better anti-
bacterial activity for the tetramic acids 1 and reasonable
correlation between enzyme inhibition and MIC effec-
tiveness in Streptococcus pneumoniae, whereas the more
potent UPPS inhibitors 4 are typically weaker antibacte-
rial agents. This might be attributed to differences in
physical–chemical properties and cell permeability be-
tween the two compound series.
In conclusion, new UPPS inhibitors were designed in-
spired by the binding mode of the natural substrate far-
nesyl pyrophosphate and subsequently synthesized. This
design principle may find use in the development of new
antibacterial compounds. Some of these compounds
(e.g., 1i, j, 4a) show sub-micromolar spUPPS enzyme inhi-
bition and antibacterial activity against Gram-positive
bacteria. However, in the presence of blood serum a large
shift to higher MICs in the tested Gram-positive bacteria
is observed. This finding will require further optimization
of compound properties to develop the UPPS inhibitors
shown here into new antibacterial agents for in vivo use.
10. Rondeau, J.-M.; Bitsch, F.; Bourgier, E.; Geiser, M.;
Hemmig, R.; Kroemer, M.; Lehmann, S.; Ramage, P.;
Rieffel, S.; Strauss, A.; Green, J. R.; Jahnke, W. Chem-
MedChem 2006, 1, 267.