Inhibitors of the acetyltransferase domain of N-acetylglucosamine-1- phosphate-uridylyltransferase/glucosamine-1-phosphate-acetyltransferase (GlmU). Part 2: Optimization of physical properties leading to antibacterial aryl sulfonamides

Article abstract of DOI:10.1016/j.bmcl.2012.10.003

A previously described aryl sulfonamide series, originally found through HTS, targets GlmU, a bifunctional essential enzyme involved in bacterial cell wall synthesis. Using structure-guided design, the potency of enzyme inhibition was increased in multiple isozymes from different bacterial species. Unsuitable physical properties (low Log D and high molecular weight) of those compounds prevented them from entering the cytoplasm of bacteria and inhibiting cell growth. Further modifications described herein led to compounds that possessed antibacterial activity, which was shown to occur through inhibition of GlmU. The left-hand side amide and the right-hand side sulfonamides were modified such that enzyme inhibitory activity was maintained (IC₅₀ <0.1 μM against GlmU isozymes from Gram-negative organisms), and the lipophilicity was increased giving compounds with Log D -1 to 3. Antibacterial activity in an efflux-pump deficient mutant of Haemophilus influenzae resulted for compounds such as 13.

Full text of DOI:10.1016/j.bmcl.2012.10.003

Bioorganic & Medicinal Chemistry Letters 22 (2012) 7019–7023
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Inhibitors of the acetyltransferase domain of N-acetylglucosamine-1-phos-
phate-uridylyltransferase/glucosamine-1-phosphate-acetyltransferase (GlmU).
Part 2: Optimization of physical properties
leading to antibacterial aryl sulfonamides
Suzanne S. Stokes a, , Robert Albert a, Ed T. Buurman a, Beth Andrews a, Adam B. Shapiro a,
⇑
Oluyinka M. Green a, Andrew R. McKenzie a, Ludovic R. Otterbein b
a Infection Innovative Medicines Unit, AstraZeneca R&D Boston, 35 Gatehouse Drive, Waltham, MA 02451, USA
b Discovery Sciences, AstraZeneca R&D Alderley Park, Macclesfield, Cheshire SK 10 4TG, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A previously described aryl sulfonamide series, originally found through HTS, targets GlmU, a bifunc-
tional essential enzyme involved in bacterial cell wall synthesis. Using structure-guided design, the
potency of enzyme inhibition was increased in multiple isozymes from different bacterial species.
Unsuitable physical properties (low LogD and high molecular weight) of those compounds prevented
them from entering the cytoplasm of bacteria and inhibiting cell growth. Further modifications described
herein led to compounds that possessed antibacterial activity, which was shown to occur through inhi-
bition of GlmU. The left-hand side amide and the right-hand side sulfonamides were modified such that
Received 15 June 2012
Revised 24 September 2012
Accepted 1 October 2012
Available online 11 October 2012
Keywords:
Antibacterial
Cell wall biosynthesis
N-Acetylglucosamine-1-phosphate
uridylyltransferase
Aryl sulfonamide
enzyme inhibitory activity was maintained (IC50 <0.1 lM against GlmU isozymes from Gram-negative
organisms), and the lipophilicity was increased giving compounds with LogD À1 to 3. Antibacterial activ-
ity in an efflux-pump deficient mutant of Haemophilus influenzae resulted for compounds such as 13.
Ó 2012 Elsevier Ltd. All rights reserved.
Physical properties
There is an increasing need for new agents to combat patho-
genic bacteria. The rate of resistance to the current arsenal of anti-
bacterial compounds is rising, and there has been a concomitant
decrease in the number of new agents approved to treat these
problematic bacteria.1–3 A number of strategies are being pursued
to discover new antibacterial compounds with activity against
resistant organisms, including making improvements on existing
chemical classes or finding new chemical scaffolds acting at either
known or novel antibacterial targets.4,5 One advantage to novel
scaffolds which act on novel antibacterial targets is that pre-
existing resistance is not expected and therefore the drug should
retain its efficacy even against clinically drug-resistant bacterial
strains. One such novel target is N-acetylglucosamine-1-phosphate
uridylyltransferase (GlmU).
GlmU is essential in both Gram-positive and Gram-negative
bacteria. It performs key steps in the synthesis of UDP-N-acetylglu-
cosamine, an intermediate in the peptidoglycan and lipopolysac-
charide biosynthetic pathways. It is a bifunctional enzyme in
which the carboxyl terminal domain possesses acetyltransferase
activity and the amino terminal domain has uridylyltransferase
activity.6–8 The acetyltransferase activity of GlmU was validated
in vitro as an antibacterial target in Haemophilus influenzae
(H. influenzae).9 GlmU does not have a high homology with eukary-
otic acetyltransferase or uridylyltransferase enzymes. Therefore,
inhibitors of this enzyme should be selective for the bacterial
enzymes.
Previous reports of inhibitors of GlmU include compounds act-
ing at both the acetyltransferase and uridylyltransferase domains.
Inhibitors of the acetyltransferase activity include 2-phen-
ylbenzofurans formed by fermentation,10 arylamines found by a
novel, specific assay,11 and a sugar analog that was reported to
be a weak inhibitor of the Mycobacterium tuberculosis GlmU iso-
zyme.12 A series of aminopiperidines inhibited uridylyltransferase
activity allosterically.13 There was also a recent report of a QSAR
model applicable for finding novel GlmU inhibitors.14
AstraZeneca reported a series of arylsulfonamide inhibitors of
GlmU exemplified by 1 (Fig. 1).15 This series was identified using
high throughput screening (HTS) of AstraZeneca’s compound col-
lection in an Escherichia coli (E. coli) GlmU enzyme assay designed
to discover inhibitors of the acetyltransferase activity. X-ray
co-crystal structures from both E. coli and Streptococcus pneumo-
niae (S. pneumoniae) were used to guide compound design. The
initial hits had desirable qualities of low molecular weight
⇑
Corresponding author.
E-mail address: Sue_Stokes@vrtx.com (S.S. Stokes).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.10.003

7020
S. S. Stokes et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7019–7023
O
O
HO
O
N
O
O
S
O
O
S
O
O
S
O
O
O
H
N
O
O
HO
HO
N
H
N
H
N
H
N
N
O
1
O
2
O
O
O
O
3
MW
LogD7.4 (clogD7.4
)
IC50 (H. influenzae, S. pneumoniae), µM
1
2
3
390.5
462.5
635.7
1.95
-0.87
1.5, >200
0.6, 20
<-1 (-4.2)
<0.45, 0.47
Figure 1. Initial optimization of binding of 1 resulted in 3, which has improved enzyme inhibitory potency (IC50) but unsuitable low LogD and high molecular weight (Ref. 15).
Table 1
Enzyme inhibition (IC50) and antibacterial (MIC) potencies for 4–18. Measured or calculated values for LogD are also shown
O
O
S
R1
O
N
H
O
R2
Compd R1
R2
Eco GlmU IC50
Hin GlmU IC50
M)
Spn GlmU IC50
M)
Hina MIC (
ml)
lg/
LogD pH=7.4
(lM)
(
l
(
l
(ACDLogD7.4
)
O
O
O
*
N
*
*
*
HO
4
5
0.04
0.04
0.91
0.32
25.5
48.0
>64
32
À0.80
O
O
O
*
N
HO
HO
0.06
*
N
6
7
0.02
0.04
0.40
0.83
40.2
8.4
>64
32
(À0.22)
NH2
O
*
N
*
HO
HO
(1.45)
O
O
O
O
*
N
*
8
0.01
0.09
0.03
0.01
0.02
0.37
0.08
0.02
14.6
36.9
55.0
21.5
32
À0.53
À0.31
(À0.54)
À0.71
O
O
O
O
*
N
*
*
HO
HO
9
>64
32
S
*
N
10
11
O
*
N
O
*
HO
64
O
O
H
N
*
N
N
N
12
0.43
0.60
24.7
>64
À0.34
*
N
O

S. S. Stokes et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7019–7023
7021
Table 1 (continued)
Compd R1
R2
Eco GlmU IC50
M)
Hin GlmU IC50
M)
Spn GlmU IC50
M)
Hina MIC (
ml)
lg/
LogD pH=7.4
(ACDLogD7.4)
(
l
(
l
(
l
*
N
HO
*
a
a
13
0.03
0.40
0.01
0.51
5.3
32
0.06
O
O
O
a = relative absolute
*
N
HO
*
a
a
14
5.9
128
À0.39
O
O
a = relative absolute
*
N
O
N
15
16
17
18
0.06
0.80
0.09
0.68
0.03
0.52
0.03
0.67
32.0
9.0
64
(3.04)
2.49
*
O
*
N
O
O
N
N
>128
>128
>128
F
F
*
*
*
N
9.9
3.37
O
*
N
N
O
10.0
(2.04)
CN
*
a
H. influenzae acrB-deficient mutant strain.
(MW <500) and low lipophilicity (À3 < LogD < 3). By modifying the
moieties on both the left-hand and right-hand sides of the central
dimethoxyphenyl core, the potency of enzyme inhibition, as mea-
sure by assays described in Ref. 15 and reported as IC50 values, was
increased. To maximize interactions with the target protein, addi-
tional acidic groups were added such as the succinyl amide on the
left-hand side (2), and finally an additional acid on the right-hand
side, resulting in the most potent GlmU inhibitor reported, 3
(Fig. 1).
Compound 3 proved GlmU isozymes from both Gram-negative
and Gram-positive species could be inhibited at sub-micromolar
IC50 values. However, during the course of this initial work to opti-
mize binding, the desirable physical properties of the hits, low
MW and favorable LogD, were lost. The compounds with these ex-
tended right-hand side sulfonamides interacting with Lys445 had
LogD ꢀ À1 (calculated LogD < À4), presumably too low to allow
passive permeationthroughthe cytoplasmic membrane.16 Although
prokaryotic and eukaryotic membranes differ in their physiology,
measurement of permeability through eukaryotic membranes has
been used as a rough surrogate for approximating the permeability
of compounds through bacterial membranes. Preliminary data from
MDCK cell line permeability measurements (data not shown) were
in agreement with the hypothesis that compounds with very low
LogD (ꢀÀ1) would not traverse the membranes. These compounds
lacked antibacterial activity and were not pursued further.
Described herein is subsequent work focused on optimization of
GlmU inhibitors with drug-like physical properties associated with
permeability through bacterial membranes. The goal was to make
potent enzyme inhibitors (IC50 <1 lM) with LogD 0–3 and MW
<550, resulting in measurable antibacterial activity.
Further optimization of left-hand side amides and right-hand
side sulfonamides led to a number of compounds with increased
enzyme potency. To reduce the overall polarity of the molecules,
side chains with greater hydrophobic character were explored
(Table 1). Compounds were made in an analogous fashion to that
previously described, as shown in Scheme 1. Compounds 13 and
14 with a cyclopropanated succinyl amide (diastereomeric mix-
ture) on the left-hand side were potent enzyme inhibitors for
E. coli and H. influenzae GlmU, but had lower potency with S. pneu-
moniae GlmU. The nanomolar potency against H. influenzae GlmU
combined with reasonable hydrophobicity, LogD À0.39, resulted
in moderate antibacterial activity against an efflux-pump deficient
mutant of H. influenzae for 13. A series of left-hand side pyridyl
substituents were also evaluated (15–18). The 2-cyano- and
2-fluoropyridyl side chains resulted in analogs with activity
against GlmU from E. coli, H. influenzae, and S. pneumoniae and
drug-like lipophilicity. However, these compounds were not
potent enough enzyme inhibitors to display antibacterial activity
at concentrations <128 lg/ml.
Modification of the right-hand side sulfonamide side chain with
hydrophobic substituents resulted in 4–11, 13, 15, and 17 (Table 1).
The diphenylamine-containing 4 exhibited high potency of enzyme
O
O
O
O
O
O
O
S
O
S
a, b
c
d, e
O
O
R1
R2
R2
N
H
SO2Cl
N
H
H2N
N
H
O
O
O
O
Scheme 1. General reaction sequence to make 1–18. For details, see Ref. 15 Reagents and conditions: (a) Ac2O pyridine, 4 h, rt; (b) ClSO3H, 48 h rt; (c) NH-R2, DIEA, MeCN,
60 °C; (d) 1 M HCl, EtOH, 80 °C; (e) R1-COOH, HATU, DIEA; if R1 contains ester, final step is LiOH, THF/H2O, 80 °C, 6 h.

7022
S. S. Stokes et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7019–7023
Table 2
Enzyme inhibition (IC50) and antibacterial (MIC) potencies for 19–21. LogD (measured or calculated) are also shown
O
OH
R1
O
N
H
S
O
R2
Compd R1
R2
GlmU Eco IC50
M)
GlmU Hin IC50
M)
GlmU Spn IC50
M)
Hina MIC (
ml)
lg/
LogD pH=7.4
(
l
(
l
(
l
(ACDLogD7.4
)
O
*
N
*
HO
19
0.05
0.06
0.28
0.03
46.6
30.3
32
32
À0.27
O
*
N
HO
*
a
a
20
21
(À0.92)
(À1.25)
O
O
O
a = relative absolute
*
N
*
HO
a
a
<0.48
0.38
7.84
128
O
O
a = relative absolute
a
H. influenzae acrB-deficient mutant strain.
inhibition and lipophilicity in a range associated with membrane
permeability. Therefore, other substituted diphenylamino-containing
analogs (5–7), as well as conformationally constrained analogs
(8–10), were made. Incorporation of the phenoxazine side chain
resulted in 8, which possessed moderate lipophilicity (LogD
À0.53) and high potencies of enzyme inhibition with the GlmU iso-
zymes from Gram-negative pathogens, E. coli and H. influenzae.
Antibacterial activity against the efflux pump-deficient H. influen-
engineered to over-express GlmU, the MICs were elevated com-
pared to the parent efflux pump-deficient mutant. Lastly, resistant
mutants were isolated and characterized which showed mutations
near the inhibitor binding site (W449G and T420I, E. coli number-
ing). All these data indicate the antibacterial activities of 8 and 19
are indeed due to inhibition of GlmU.
In summary, a sulfonamide series of GlmU inhibitors of the ace-
tyltransferase activity was initially discovered by HTS, and binding
was optimized using structure-guided design. However, these
early compounds were devoid of antibacterial activity as a conse-
quence of physical properties unsuitable for bacterial cytoplasmic
membrane penetration. Therefore, efforts were undertaken to im-
prove potency with compounds likely to have physical properties
leading to bacterial membrane permeability. Compounds with
antibacterial activity demonstrated to be through inhibition of
GlmU were made by focusing on modifications of the left-hand
side amide and right-hand side sulfonamide with less polar sub-
stituents. The compounds described in this letter do not have activ-
ity against wild-type bacterial pathogens, and therefore are still far
zae mutant strain was 32 lg/ml. Other favorable side chains in-
cluded the modified phenoxazines 9 and 10.
Combinations of the favorable substituents on the left- and right-
hand sides resulted in analogs with improved potencies of enzyme
inhibition and antibacterial activity against the H. influenzae efflux
pump-deficient mutant (Table 1: 13, 15, 17). Compound 13 with
the cyclopropanated succinyl amide and phenoxazine sulfonamide
achieved the most balanced overall profile: IC50 values of 28 nM,
9 nM, and 5.3 lM against the E. coli, H. influenzae, and S. pneumoniae
GlmU isozymes respectively, LogD 0.06, and MIC = 32
lg/ml against
the H. influenzae efflux pump-deficient mutant.
Most changes to the dimethoxyphenyl core resulted in reduced
enzyme potency (data not shown). However, removal of the methyl
group from the 4-methoxy substituent (Table 2) either results in in-
creased potency of enzyme inhibition as for 19, or was neutral with
respect to potency (20–21). One possible explanation is that the
methoxy group was pointing into a hydrophilic pocket ( Fig. 2),
and therefore by removing the methyl group, the unfavorable inter-
action was eliminated. The overall lipophilicity was actually in-
creased for monomethoxy 19 compared to dimethoxy 2 (LogD
À0.27 compared to À0.87), perhaps due to an internal hydrogen
bond between the phenolic hydrogen and sulfonyl oxygen.
Compound 19 was made using the reaction sequence shown in
Scheme 2. The methyl ester precursor that was made using the
reaction sequence described in Scheme 1 was demethylated using
LiI; concomitant formation of the phthalimide occurred under forc-
ing conditions. Hydrolysis and ring opening to the succinic acid re-
sulted in the final monomethylated analog 19. Compounds 20 and
21 were made in an analogous fashion.
The mode-of-action of the antibacterial activity of 8 and 19 was
determined using a number of methods. In a radio-labeled biosyn-
thetic precursor incorporation assay, cell-wall biosynthesis and li-
pid biosynthesis pathways were inhibited, as would be expected
when the GlmU reactions are blocked. Also, in a H. influenzae strain
Figure 2. Molecular modeling showing 19 (stick representation) docked into S.
pneumoniae GlmU (space filling representation). Phenolic hydroxyl binds in polar
binding pocket formed by backbone NH and C@O groups. Residues are colored by
atom type; hydrogen bonds are shown as dotted lines.

S. S. Stokes et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7019–7023
7023
OH
O
O
O
N
O
LiI, pyr
O
O
reflux, 4 days
O
O
S
S
N
H
50 %
N
N
O
O
O
O
OH
O
O
1N LiOH
O
55 oC, 3-5h
HO
S
N
N
H
20%
O
O
19
Scheme 2. Synthesis of monomethylated analog 19.
3. Gilbert, D. N.; Guidos, R. J.; Boucher, H. W.; Talbot, G. H.; Spellberg, B.;
Edwards, J. E.; Scheld, W. M.; Bradley, J. S.; Bartlett, J. G. Clin. Infect. Dis. 2010,
50, 1081.
4. Theuretzbacher, U. Curr. Opin. Pharmacol. 2011, 11, 433.
5. Silver, L. L. Clin. Microbiol. Rev. 2011, 24, 71.
from being antibacterial drugs. However, this work provides fur-
ther evidence that GlmU represents a useful target for new anti-
bacterial drug discovery efforts.
6. Mengin-Lecreaulx, D.; van Heigenoort, J. J. Bacteriol. 1994, 176, 5788.
7. Gehring, A. M.; Lees, W. J.; Mindiola, D. J.; Walsh, C. T.; Brown, E. D. Biochemistry
1996, 35, 579.
Acknowledgments
8. Pompeo, F.; Bourne, Y.; van Heijenoort, J.; Fassy, F.; Mengin-Lecreulx, D. J. Biol.
Chem. 2001, 276, 3833.
9. Buurman, E. T.; Andrews, B.; Gao, N.; Hu, J.; Keating, T. A.; Lahiri, S.;
Otterbein, L. R.; Patten, A. D.; Stokes, S. S.; Shapiro, A. B. J. Biol. Chem. 2011,
286, 40734.
10. L. Vertesy, M. Kurz, A. Markus-Erb, L. Toti. U.S. Patent 0,137,254, 2005.
11. Pereira, M. P.; Blanchard, J. E.; Murphy, C.; Roderick, S.; Brown, E. D. Antimicrob.
Agents Chemother. 2009, 53, 2306.
12. Li, Y.; Zhou, Y.; Mac, Y.; Li, X. Carbohydrate Res. 2011, 346, 1714.
13. Mochalkin, I.; Lightle, S.; Narasimhan, L.; Bornemeier, D.; Melnick, M.;
Vanderroest, S.; McDowell, L. Protein Sci. 2008, 17, 577.
14. Singla, D.; Anurag, M.; Dash, D.; Raghava, G. P. S. BMC Pharmacol. 2011, 11,
1471.
The authors thank the following groups: the Biosciences
Department at Infection Innovative Medicines, AstraZeneca, for
determination of IC50 and MIC values; Discovery Sciences of Astra-
Zeneca Alderley Park and the Structural Biology Department of
AstraZeneca Boston for solving enzyme-inhibitor X-ray structures;
and the Drug Metabolism & Pharmacokinetics Department at
AstraZeneca Boston for physical properties determinations. The
authors also thank Haihong Ni and John Manchester for computa-
tional chemistry support.
15. Green, O. M.; McKenzie, A. R.; Shapiro, A. B.; Otterbein, L.; Ni, H.; Patten, A.;
Stokes, S.; Albert, R.; Kawatkar, S.; Breed, J. Bioorg. Med. Chem. Lett. 2012, 22,
1510.
References and notes
16. For all LogD values reported: LogD 7.4 was measured using a miniaturized
shake-flask octanol:water assay with HPLC-UV quantitation at pH 7.4, cLogD
7.4 values were calculated using software from Advanced Chemistry
Development Labs.
1. Talbot, G. H.; Bradley, J.; Edwards, J. E.; Gilbert, D.; Scheld, M.; Bartlett, J. G. Clin.
Infect. Dis. 2006, 42, 657.
2. Spellberg, B.; Guidos, R.; Gilbert, D.; Bradley, J.; Boucher, H. W.; Scheld, W. M.;
Bartlett, J. G.; Edwards, J. E. Clin. Infect. Dis. 2008, 46, 155.