Synthesis and structural-activity relationships of 3-hydroxyquinazoline-2, 4-dione antibacterial agents

Article abstract of DOI:10.1016/j.bmcl.2004.06.063

The synthesis and SAR of the 3-hydroxyquinazoline diones, a novel class of bacterial topoisomerase inhibitors, is reported. A series of 3-hydroxyquinazoline-2,4-diones was synthesized and evaluated for antibacterial activity. This series represents a novel addition to the DNA gyrase inhibitor class of antibacterials. Appropriate substitutions onto the core template yielded compounds with excellent potency against E. coli gyrase and significant in vitro Gram-negative and Gram-positive antibacterial activity.

Full text of DOI:10.1016/j.bmcl.2004.06.063

Bioorganic & Medicinal Chemistry Letters 14 (2004) 4405–4409
Synthesis and structural–activity relationships of
3-hydroxyquinazoline-2,4-dione antibacterial agents
a
a
a
Tuan P. Tran,a, Edmund L. Ellsworth, Michael A. Stier, John M. Domagala,
H. D. Hollis Showalter,a Stephen J. Gracheck,b Martin A. Shapiro,b
Themis E. Joannidesb and Rajeshwar Singhc
*
aDepartment of Chemistry, Pfizer Global Research and Development, Ann Arbor Laboratories,
2800 Plymouth Road, Ann Arbor, MI 48105, USA
bDepartment of Antibacterial Pharmacology, Pfizer Global Research and Development, Ann Arbor Laboratories,
2800 Plymouth Road, Ann Arbor, MI 48105, USA
cNAEJA Pharmaceuticals, 4290-91A Street, Edmonton, Alta., Canada T6E 5V2
Received 15 April 2004; revised 21 June 2004; accepted 21 June 2004
Abstract—A series of 3-hydroxyquinazoline-2,4-diones was synthesized and evaluated for antibacterial activity. This series repre-
sents a novel addition to the DNA gyrase inhibitor class of antibacterials. Appropriate substitutions onto the core template yielded
compounds with excellent potency against E. coli gyrase and significant in vitro Gram-negative and Gram-positive antibacterial
activity.
Ó 2004 Elsevier Ltd. All rights reserved.
The fluoroquinolones (1) constitute a major class of
antibacterial agents with both Gram-positive and
Gram-negative activity. The antibacterial activity of this
series results from selective inhibition against type-2
topoisomerase enzymes essential for DNA replication,
namely gyrase and topoisomerase IV.1–3 Due to their
increased use over the past decades as one of the premier
classes of antibacterial agents, resistance is developing
toward the fluoroquinolones and this is contributing
to an emerging world-wide health problem within
the infectious diseases therapeutic area.4 This trend,
along with safety issues associated with the fluoroquino-
lones,5–10 continues to provide the impetus for the
discovery of improved agents within this mechanistic
class.
As part of a program to discover new classes of antibac-
terial agents, we identified the 3-hydroxyisoquinolones
(2) as possessing moderate in vitro antibacterial activ-
ity.11 Further pursuit in this area led to the discovery
of the 3-hydroxyquinazoline-2,4-diones (3). This series
incorporates structural features imbedded in both the
well-known fluoroquinolones and the hydroxyisoquino-
lones (Scheme 1).
O
O
O
O
F
OH
O
N
F
OH
O
F
N
OH
R7
N
R7
N
R7
R8 R1
R8 R1
R8
1
3
2
Scheme 1.
Keywords: 3-Hydroxyquinazoline-2,4-diones; Antibacterial agents; Bacterial topoisomerase inhibitors.
*
Corresponding author. Tel.: +1-734-622-5496; fax: +1-734-622-2265; e-mail: tuan.tran@pfizer.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.06.063

4406
T. P. Tran et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4405–4409
One important difference is the presence of the 3-hy-
droxamic acid instead of the typical 3-carboxylic acid
moiety in the fluoroquinolones. The carboxylic acid
has long been considered essential for antibacterial
activity in the fluoroquinolone field.12,13 The acidity of
the 3-hydroxyl of the hydroxyquinazoline-2,4-diones
(pKa =6.3–7.0) was found to be similar to that of the
3-carboxylic functionality of the fluoroquinolones
(pKa =5.6–6.4) and thus has turned out to be an appro-
priate substitution. We also have shown that many of
the hydroxyquinazoline-2,4-diones inhibited bacterial
DNA gyrase, the very same enzyme targeted by the
fluoroquinolones (Table 1). Because of the important
features shared by the two series, our strategy was to
apply the vast knowledge of the fluoroquinolone field
toward optimizing antibacterial activity within the
hydroxyquinazoline-2,4-diones, while at the same time
attempting to find new structure–activity relationship
(SAR) space that might alleviate some of the known
toxicities associated with the fluoroquinolones. Toward
Table 1. In vitro antibacterial and gyrase activity
Compd
R1
R7
R8
Minimum inhibitory concentrations (MIC), lg/mL
Gram-negative organisms Gram-positive organisms
E. coli WT E. coli LKY E. coli TLC E. faec S. aureus S. pyog
Gyrase-drug induced
cleavage, lM;
E. coli
Ciprofloxacin
0.06
64
0.06
32
0.06
16
0.4
NT
4
8
2
0.4
10a
H
N
N
N
N
N
N
H
H
H
H
H
H
32
>100
10b
10c
10d
10e
10f
Me
Et
32
>64
8
16
2
8
>64
NT
NT
64
>64
32
1
>64
32
16
8
>100
>100
7
1
0.1
0.5
32
0.13
0.5
32
8
13
>64
16
>64
16
32
>100
10g
H
>64
32
32
64
64
32
>100
N
F
F
H2N
H2N
10h
10i
H
F
1
8
0.3
0.25
NT
16
32
32
4
2
N
N
2
1
1
16
0.28
H2N
H2N
10j
Cl
2
0.5
32
16
4
2
1
N
10k
10l
H
H
8
8
1
1
NT
4
16
1
NT
6.4
N
H
0.125
0.125
0.5
H2N
N
N
H2N
10m
10n
10o
H
H
H
8
32
2
4
1
0.5
2
16
16
16
16
16
4
32
2
2.2
6.2
8
H
N
H
N
F3C
N
H
>64
0.5
16
N
H
10p
10q
H2N
N
H
H
1
8
0.25
2
0.25
1
4
4
2
0.7
5
H
H
N
H
H
64
32
32
N
10r
N
N
H
H
H
4
16
2
1
32
NT
64
32
>64
64
32
>64
64
7.8
18
10s
H2N
N
N
0.25
8
0.5
2
N
10t
>64
0.4
NT=Not tested.

T. P. Tran et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4405–4409
4407
realizing the first part of this goal, this paper describes
the synthesis and in vitro antibacterial activity of the
3-hydroxyquinazoline-2,4-dione series.
hydrochloride gave hydroxamate 13, which was then
cyclized with phosgene to provide the quinazoline-2,4-
dione nucleus 8. Displacement of the 7-fluorine with
heterocyclic amino side chain and then deprotection
with trifluoroacetic acid, as according to Scheme 2,
afforded targets 10 with aryl R1 substitutions.
The 3-hydroxyquinazoline-2,4-diones were prepared
according to Schemes 2–4. In most cases, target mole-
cules were derived from appropriately substituted anth-
ranilic acid precursors. Typically, compounds with
nonaryl R1 substitution were synthesized according to
Scheme 2. Acid 4 was treated with carbonyldiimidazole
in tetrahydrofuran at reflux to form the isatoic anhy-
dride 5.14 The anhydride intermediate was reacted with
O-tert-butyl hydroxylamine hydrochloride in the pres-
ence of triethylamine to afford hydroxamate 6, which
was then cyclized with phosgene to provide the quinaz-
oline-2,4-dione nucleus 7. Alkylation of 7 gave com-
pound 8 with the desired N-1 substitution (R1). The
next step involved the installation of R7 side chains,
which were typically pyrrolidines with amino functional-
ity bearing an acid labile protecting group. These
heterocyclic amine side chains were prepared according
to established literature procedures.15–18 Aromatic nu-
cleophilic displacement of the C-7 fluorine by R7 hetero-
cyclic amine in N,N-dimethylacetamide at elevated
temperature gave penultimate 9, which was treated then
with trifluoroacetic acid to remove the tert-butyl group
from the 3-hydroxyl moiety as well as the protecting
group of the R7 side chain to provide target 10.
Intermediate 8 was also synthesized via an alternate
route (Scheme 4). Lithiation of trifluorobenzoic acid
(14) followed by an electrophilic quench installed the
appropriate C-8 substituent (R8) for certain core tem-
plates. Treatment of acid 15 with a carbodiimide and
O-tert-butyl hydroxylamine hydrochloride gave hy-
droxamate 16. Hydroxamate 16 was deprotonated with
sodium hydride and the resulting anion was reacted with
an alkyl or aryl isocyanate to provide urea 17, which was
cyclized with sodium hydride to form the quinazoline-
dione core 8.
All synthesized 3-hydroxyquinazoline-2,4-diones were
tested against representative Gram-negative and
Gram-positive pathogens using standard microtitration
techniques.19 The strains reported in this paper are
E. coli MC4100 (wild type), E. coli LKY (cell wall defi-
cient, semi-permeable), E. coli TLC (membrane efflux
pump deficient), Enterococcus faecalis, Staph. aureus,
and Strep. pyogenes. E. coli LKY and E. coli TLC, lab-
oratory-constructed strains, were used to determine the
intrinsic activity for analogues in the absence of physio-
chemical factors that influence permeability and trans-
port, respectively. All compounds were less active
against the Gram-negative wild-type strain than both
the mutant strains. These results suggest that against
native E. coli, the hydroxyquinazoline-2,4-diones have
Scheme 3 shows the synthetic route of targets with aryl
R1 functional groups. Acid 11 and aniline were treated
with lithium diisopropylamide separately, then the two
species were combined to form 12. Treatment of acid
12 with a carbodiimide and O-tert-butyl hydroxylamine
O
O
O
O
F
F
O-tBu
F
F
O-tBu
O
F
F
F
F
OH
NH2
(a)
O
N
H
(c)
N
(b)
N
H
O
NH2
N
H
R8
R8
R8
N
R8
6
5
4
7
O
N
O
N
O
N
O-tBu
O
F
OH
O
F
O-tBu
O
F
F
(f)
N
N
(e)
(d)
R7
R7
R8 R1
R8 R1
R8 R1
9
10
8
Scheme 2. Reagents and conditions: (a) CDI (1.5equiv), THF, reflux, overnight; (b) t-BuONH3Cl (1.5equiv), TEA (2equiv), THF, room temp to
reflux, 16h, 52–84% (two steps); (c) COCl2 (1.9 solution in toluene, 1.2equiv), THF, reflux, 2–5h, 67–88%; (d) R1–X (1.3equiv), NaH (1.5equiv),
DMF, 0–50°C, 6h, 50–87%; (e) R7 (amino side chain) (2equiv), TEA (2–5equiv), DMA, 90°C, overnight, 22–70%; (f) TFA, CH2Cl2, overnight, 80–
90%.
O
F
O
O
O
F
F
F
F
F
F
O-tBu
F
F
O-tBu
O
OH
OH
N
H
N
(a)
(b)
(c)
NH
NH
N
R8
R8 R1
R8 R1
R8 R1
11
12
13
8
Scheme 3. Reagents and conditions: (a) LDA (3equiv), aniline (1.5equiv), À78–0°C, 4h, 55–84%; (b) t-BuONH3Cl (1.3equiv), TEA (2equiv), EDCI
(1.5equiv), CH2Cl2, overnight, 73–90%; (c) COCl2 (1.9 solution in toluene, 1.2equiv), THF, reflux, 2–5h, 68–85%.

4408
T. P. Tran et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4405–4409
O
O
F
O
F
O
F
F
O-tBu
F
F
O-tBu
F
F
F
F
N
H
N
O
(c)
OH
OH
(b)
(a)
NH
R1
F
F
R8
R8
R8
14
16
17
15
(d)
8
Scheme 4. Reagents and conditions: (a) LiHMDS (2.2equiv), THF, R8X (1.5equiv), À78–0°C, 2–4h, 48–93%; (b) t-BuONH3Cl (1.3equiv), TEA
(2equiv), EDCI (1.5equiv), CH2Cl2, overnight, 81–94%; (c) NaH (1.5equiv), RNCO (2–3equiv), CH2Cl2, 0°C to reflux, 1–6h, 67–85%; (d) NaH
(1.5equiv), DMF, 0°C to reflux, 6h, 63–96%.
limited permeability across the cell membrane and once
inside the cell they are actively pumped out. This is a
problem also encountered by a number of antibacterial
agents.20 The minimum inhibitory concentrations
(MICs) of this series are summarized in Table 1, and
compared to ciprofloxacin. Overall, this series shows
lower antibacterial activity than ciprofloxacin.
Substitution at C-8 was explored and MICs and gyrase
activity are given for R8 =H, F, Cl (10h, 10i, 10j, respec-
tively). The fluoro substituent (10i) gave the best enzyme
potency (0.28lg/mL), but this did not translate to im-
proved in vitro antibacterial activity versus the unsubsti-
tuted and the 8-Cl analogues (10h and 10j). The
unsubstituted analogue (10h) had the best Gram-nega-
tive activity, with MICs two- to fourfold lower than sub-
stituted analogues 10i and 10j.
The compounds were also tested for their inhibition of
DNA gyrase, which was isolated and purified from
E. coli H560,21 to determine the effects of these com-
pounds at the target enzyme level. An assay was
employed that measures the lowest concentration of
drug (lg/mL), which will produce a detectable level of
cleavage from relaxed bacterial Col E1 plasmid DNA,
as visualized by agarose gel elecrophoresis and staining
with ethidium bromide.21 The in vitro results of the
hydroxyquinzoline-2,4-diones showed an important cor-
relation between the ability to inhibit the DNA gyrase to
antibacterial activity. It was evident from the data pre-
sented in Table 1 that all compounds with poor gyrase
activity also displayed poor or no antibacterial activity.
But it was also notable that the best gyrase activity did
not always translate to the best in vitro activity (10i vs
10j, 10t). This trend was also observed with the fluoro-
quinolones.22
The SAR of the C-7 position was examined more exten-
sively. For comparison purposes, keeping R1 and R8
constant, many different R7 groups were synthesized
and their SAR evaluated. Based on the fluoroquinolone
literature, we chose to append many nitrogen heterocy-
cles, with functionalized pyrrolidines being the domi-
nant side chain. Two nonpyrrolidinyl substitutions
were examined. The azetidinylamine (10s) displayed
good Gram-negative activity, but not Gram-positive,
and had poor enzymatic activity. While the imidazole
analogue (10t) was an effective inhibitor of the gyrase
enzyme, this did not translate to antibacterial activity.
In general, an amino substituent appended to the 3-posi-
tion of the pyrrolidine side chain enhanced antibacterial
activity, with primary amino analogues (10k, 10l, 10m)
more active than secondary or tertiary amino com-
pounds (10n, 10o, 10q). Branched alkyl substitutions al-
pha to the distal amine gave some of the most active
compounds in this series. The mono-methyl (10l) and
the [3.1.0] bicycle (10p) substitutions both imparted
good broad-spectrum antibacterial activity and enzy-
matic activity.
The effect of substitution at the N-1, C-7, and C-8 posi-
tions (R1, R7, and R8, respectively) was examined. For
R1 =alkyl or aryl, the data shown in Table 1 indicate
that substituent size is important for both antibacterial
and gyrase activity. Both small (10a, 10b, 10c) and large
(10f, 10g) substituents did not impart good activity. The
cyclopropyl and cyclopropylmethyl substitutions (10d,
10e) showed the most promising results. They both dis-
played notably improved Gram-negative activity and
drastically better enzymatic activity. In particular, the
cyclopropyl moiety (10e) provided the broadest spec-
trum antibacterial activity and good potency toward
gyrase. These data demonstrate both similarities and
differences between the hydroxyquinazoline-2,4-diones
and the fluoroquinolones. Whereas the bulky N-1 aro-
matic substitutions maintain potency in the fluoroquino-
lones, they are not tolerated in these hydroxy-
quinazoline-2,4-diones.23 Meanwhile, the N-1 cyclopro-
pyl group, a prominent presence in the fluoroquinolone
antibiotics, is an essential substitution for this series.
In summary, we have described the discovery of the 3-
hydroxyquinazoline-2,4-dione series that possesses
in vitro antibacterial activity and good potency toward
the gyrase enzyme. The two remarkable compounds of
this series (10l, 10p) have significant broad-spectrum
antibacterial properties. While compounds of this class
display a structural similarity to the fluoroquinolone
antibacterials and inhibit the same enzymatic target,
they exhibit some markedly different SAR. Pursuing
and exploring these differences will enable us to expand
on the already well-studied SAR of the fluoroquinolo-
nes, and these findings can be exploited toward the
exploration of new antibacterial agents within this
mechanistic class.

T. P. Tran et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4405–4409
4409
13. Wentland, M. P. Structure–Activity Relationships of
Fluoroquinolones. In The New Generation of Quinolones;
Siporin, C., Heifetz, C. L., Domagala, J. M., Eds.; Marcel
Dekker: New York, 1990; pp 1–43.
References and notes
1. Drlica, K.; Zhao, X. Microbiol. Mol. Biol. Rev. 1997, 61,
377.
14. Jacobs, R. L. J. Het. Chem. 1970, 7, 1337.
15. Sanchez, J. P.; Domagala, J. M.; Heifetz, C. L.; Priebe, S.
R.; Sesnie, J. A.; Trehan, A. K. J. Med. Chem. 1992, 35,
1764.
16. Li, Q.; Chu, D. T. W.; Claiborne, A.; Cooper, C. S.; Lee,
C. M.; Raye, K.; Berst, K. B.; Donner, P.; Wang, W.;
Hasvold, L.; Fung, A.; Ma, Z.; Tufano, M.; Flamm, R.;
Shen, L. L.; Baranowski, J.; Nilius, A.; Alder, J.;
Meulbroek, J.; Marsh, K.; Crowell, D.; Hui, Y.; Seif, L.;
Melcher, L. M.; Henry, R.; Spanton, S.; Faghih, R.; Klein,
L. L.; Tanaka, S. K.; Plattner, J. J. J. Med. Chem. 1996,
39, 3070.
17. Brighty, K. E.; Castaldi, M. J. Synlett 1996, 11, 1097.
18. Domagala, J. M.; Heifetz, C. L.; Mich, T. F.;
Nichols, J. B. J. Med. Chem. 1986, 29, 445.
19. Cohen, M. A.; Griffin, T. J.; Bien, P. A.; Heifetz, C. L.;
Domagala, J. M. Antimicrob. Agents Chemother. 1985, 28,
766.
2. Hoshino, K.; Kitamura, A.; Morrissey, I.; Sato, K.; Kato,
J.; Ikeda, H. Antimicrob. Agents Chemother. 1994, 38,
2623.
3. Shen, L. L.; Chu, D. T. W. Curr. Pharm. Des. 1996, 2, 195.
4. Roth, V. R.; Jarvis, W. R. Emerging Ther. Targets 1999, 3,
73.
5. Hoshino, K.; Sato, K.; Une, T.; Osada, Y. Antimicrob.
Agents Chemother. 1989, 33, 1816.
6. Hoshino, K.; Sato, K.; Akahane, K.; Yoshida, A.;
Hayakawa, I.; Sato, M.; Une, T.; Osada, Y. Antimicrob.
Agents Chemother. 1991, 35, 309.
7. Hussy, P.; Maass, G.; Tummler, B.; Grosse, F.; Schom-
burg, U. Antimicrob. Agents Chemother. 1986, 29, 1073.
8. Moreau, N. J.; Robaux, H.; Baron, L.; Tabary, X.
Antimicrob. Agents Chemother. 1990, 34, 1955.
9. Oomori, Y.; Yasue, T.; Aoyama, H.; Hirai, K.; Suzue, S.;
Yokota, T. Antimicrob. Agents Chemother. 1988, 22,
91.
20. Kaatz, G. W.; Seo, S. M.; Ruble, C. A. Antimicrob. Agents
Chemother. 1993, 37, 1086.
21. Domagala, J. M.; Hanna, L. D.; Heifetz, C. L.; Hutt, M.
P.; Mich, T. F.; Sanchez, J. P.; Solomon, M. J. Med.
Chem. 1986, 29, 394.
10. Domagala, J. M. J. Antimicrob. Chemother. 1994, 33, 685.
11. Stier, M. A.; Watson, B. M.; Sanchez, J. P.; Domagala, J.
M.; Gracheck, S.; Joannides, E. T.; Miller, P. F.; Olson,
E.; Renau, T.; Shapiro, M.; Micetich, R. G.; Singh, R.;
Valsburg, A.; Wu, K.; Amegadzie, A. K.; Vo, D. Poster
1503. 40th ICAAC Meeting, September 17–20, Toronto,
Ontario, Canada.
22. Koga, H.; Itoh, A.; Murayama, S.; Suzue, S.; Irikura, T. J.
Med. Chem. 1980, 23, 1358.
23. Suto, M. J.; Domagala, J. M.; Roland, G. E.; Mailloux,
G. B.; Cohen, M. A. J. Med. Chem. 1992, 35, 4745.
12. Chu, D. T. W. Isothiazoloquinolones: Antibacterial and
Antineoplastic Agents. Drugs Future 1992, 17, 1101.