Discovery of pyrazolthiazoles as novel and potent inhibitors of bacterial gyrase

Article abstract of DOI:10.1016/j.bmcl.2010.03.052

Bacterial DNA gyrase is an attractive target for the investigation of new antibacterial agents. Inhibitors of the GyrB subunit, which contains the ATP-binding site, are described in this communication. Novel, substituted 5-(1H-pyrazol-3-yl)thiazole compounds were identified as inhibitors of bacterial gyrase. Structure-guided optimization led to greater enzymatic potency and moderate antibacterial potency. Data are presented for the demonstration of selective enzyme inhibition of Escherichia coli GyrB over Staphlococcus aureus GyrB.

Full text of DOI:10.1016/j.bmcl.2010.03.052

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2828–2831
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of pyrazolthiazoles as novel and potent inhibitors of bacterial gyrase
a
b
a
a
Steven M. Ronkin a, , Michael Badia , Steve Bellon , Anne-Laure Grillot , Christian H. Gross ,
Trudy H. Grossman c, Nagraj Mani a, Jonathan D. Parsons a, Dean Stamos a, Martin Trudeau a, Yunyi Wei a,
Paul S. Charifson a
*
a Vertex Pharmaceuticals Inc., 130 Waverly Street, Cambridge, MA 02139, USA
b Constellation Pharmaceuticals, 148 Sydney St., Cambridge, MA 02139, USA
c Tetraphase Parmaceuticals, Inc, 480 Arsenal St., Suite 110, Watertown, MA 02472, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Bacterial DNA gyrase is an attractive target for the investigation of new antibacterial agents. Inhibitors of
the GyrB subunit, which contains the ATP-binding site, are described in this communication. Novel,
substituted 5-(1H-pyrazol-3-yl)thiazole compounds were identified as inhibitors of bacterial gyrase.
Structure-guided optimization led to greater enzymatic potency and moderate antibacterial potency.
Data are presented for the demonstration of selective enzyme inhibition of Escherichia coli GyrB over
Staphlococcus aureus GyrB.
Received 30 November 2009
Revised 4 March 2010
Accepted 5 March 2010
Available online 15 March 2010
Keywords:
Bacterial gyrase
Antibacterials
Ó 2010 Elsevier Ltd. All rights reserved.
Due to the emergence of bacteria resistant to current therapeu-
tic agents the investigation of new antibiotics has become a neces-
sary endeavor.1 Bacterial DNA gyrase is a well-established target
with commercial success, exemplified by quinolones such as cipro-
floxacin (1a).2 However, resistance is now a problem for this class
of antibacterials in addition to most other classes.3 Gyrase consists
of two heterodimeric subunits, GyrA and GyrB. The quinolone class
of molecules inhibits GyrA and induces cell death by trapping the
gyrase–DNA complex, inducing oxidative damage, and preventing
DNA replication.4 Compounds like novobiocin (1b) inhibit GyrB,
which blocks ATPase activity, thus depriving the source of energy
needed for DNA replication.5 GyrB as a target offers an opportunity
such as a lack of cross-resistance to the quinolones. Some other
known GyrB inhibitors are the cyclothialidines (represented by
1c) and pyrrolamides (1d) from Astra-Zeneca.6,7 We have previ-
ously reported the aminobenzimidazoles8 (1e) and here we de-
scribe the pyrazolthiazole class of GyrB inhibitors (1) (Fig. 1).
In our efforts to identify novel starting points to discover inhib-
itors of the Gyrase B (GyrB) subunit, a high-throughput screen de-
signed to identify compounds inhibiting the ATP-ase activity of the
GyrB subunit was conducted. Compound 1 was identified and was
shown to be a moderately potent inhibitor of GyrB (E. coli
molecule (Fig. 2). These hydrogen bonds correspond with those
of the terminal carbamate of novobiocin in the published E. coli/
novobiocin structure.8 Initial investigation of the 5-position of
the pyrazole demonstrated that an ethyl ester was equipotent to
the methylthiol. Direct alkyl analogs such as the ethyl (1g), were
less potent (E. coli Ki = 4.8
l
M) (Table 1).
NH2
O
N
O
OH
O
O
O
OH
OH
O
F
H
N
OH
O
N
O
N
O
O
HN
1b
1a
Cl
Cl
HN COOH
O
H
O
N
OH S
O
OH
NH2
N
H
O
HN
O
O
O
N
O
OH
N
HO
N
H
1d
HO
1c
Ki = 2.2 lM). An X-ray crystal structure of compound 1 bound to
F
Staphlococcus aureus GyrB was obtained, and showed a hydrogen
bond network between the pyrazole moiety of 1, Asp81 (Asp73
in E. coli numbering) and a highly conserved structural water
N
HN N
S
O HN
S
S
N
N
1
N
N
H
N
H
1e
* Corresponding author. Tel.: +1 617 444 6507; fax: +1 617 444 7822.
E-mail address: steven_Ronkin@vrtx.com (S.M. Ronkin).
Fig. 1. Inhibitors of bacterial DNA gyrase
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.03.052

S. M. Ronkin et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2828–2831
2829
pyrazole was warranted. Since it appeared that substituents of
the size of carboethoxy could be tolerated, amides and carbamates
were explored. The ethyl amides at the 5-position of the pyrazole
shown in Table 3 were observed to be approximately twice as po-
tent GyrB inhibitors as the corresponding ethyl ester analogs (com-
pounds 3c vs 2k and 3g vs 2i). Some of these analogs exhibited
moderate antibacterial activity. The nitrogen linked piperidine
substituted at the 4-thiazole position, 3b, was potent against the
enzyme and showed that it could penetrate the bacteria, demon-
strating weak MIC against the E. coli tol C pump knockout strain
(MIC = 32 l/mL). However, the N-linked piperidine compounds
(3b and other substituted N-linked piperidines not shown) were
also found to possess chemical stability issues limiting their useful-
ness. The (S)-2-piperidine 3d demonstrated some S. aureus and
E. coli tol C antibacterial potency with significant S. pneumoniae
activity. The R-isomer 3e was very weak showing only modest
binding relative to the S-enantiomer. A more lipophilic substituent,
such as the phenyl 3c, showed greater antibacterial potency. These
compounds however, lacked sufficient solubility to be investigated
further. The compounds were tested intermittently for S. aureus
enzyme inhibitory gyrase activity and generally were equivalent
with the E. coli Ki values in the ester and amide series.
The carbamates shown in Table 4 were extremely potent in
terms of enzymatic inhibition, due to formation of an additional
hydrogen bond between the sidechain of Asn 46 and the carbamate
carbonyl, and demonstrated the most potent E. coli tol C activity.
These carbamates however, appeared to be effluxed from the bac-
teria since there was no observable antibacterial potency against
wild type E. coli strains. These compounds showed moderate inhi-
bition of S. aureus gyrase but antibacterial activity was not ob-
served against the S. aureus ATCC29213 and S. pneumoniae
Fig. 2. Crystallographic data showing the pyrazole moiety of compound 1 binding
interaction with Asp81 and a conserved water molecule in S. aureus GyrB. PDE code
3G75.
Table 1
Escherichia coli enzyme activity
Y
N
NH
N
S
Ar
ATCC10015 strains at MIC concentrations of 64 l/mL. While the
piperidine analogs (4b and 4c) penetrated the bacteria, the cyclo-
hexyl compound (4e) did not and was inactive, though potent
against the enzyme. Interestingly, the S. aureus enzyme activity
was significantly weaker. From the X-ray structure (Fig. 3A and
B), it was observed that there was less space in this portion of
the binding site of the S. aureus GyrB subunit relative to E. coli.
The shape of the binding pocket formed by Ile-51, Leu-103, and
Ile-175 in the observed S. aureus crystal structure, containing the
methyl carbamate analog 4f, is significantly different from the ob-
served E. coli GyrB structure containing a propargyl carbamate ana-
log 4b (Fig. 3A). In the E. coli enzyme structure, the isoleucines are
valines and the Leu-103 is a methionine (Fig. 3A). The propargyl
carbamate when shown superimposed on the S. aureus GyrB sur-
face revealed a lack of space for efficient binding (Fig. 3B), which
Compound
Ar
Y
E. coli Ki (lM)
1
Thiophene
Thiophene
Thiophene
Phenyl
SMe
2.9
2.6
4.8
3.5
1.7
1f
1g
1h
1i
COOEt
CH2CH3
SMe
Phenyl
COOEt
To optimize the substituent at the 2-position of the thiazole,
examination of the structure suggested that targeting Arg 136
(E. coli numbering) as an additional residue for hydrogen bond for-
mation could lead to additional potency. To that effect, the 2-(pyri-
din-3-yl)thiazole pyrazole compound 2e was prepared. Compound
2e shows a 5–10-fold improvement in potency over compounds 1f
and 1i with a Ki of 0.33 lM The hydrogen of the pyridyl nitrogen
makes a hydrogen bond with Arg 136 (data not shown), which
mimics the hydrogen bond observed with the exocyclic carbonyl
oxygen of the coumarin ring in the published novobiocin struc-
ture.8 From the X-ray structure of compound 1, it was also deter-
mined that substituents bigger than methyl at the 4-position of
the thiazole ring should be tolerated (compounds 2a–2d, 2f–2k).
It was found that this position of the scaffold preferred lipophilic
substituents, such as cyclohexyl 2i (Ki of 0.12
potent than cyclopropyl 2j (Ki of 0.24 M), and 2.5-fold more po-
tent than methyl 2e (Ki of 0.33 M). The 4-position of the thiazole
lM) which was more
l
l
could also tolerate the introduction of heteroatoms (2c, 2d, 2f–h).
This resulted in compounds that maintained potency but had a
better solubility profile. However, antibacterial activity was not
observed against wild type strains of E. coli, S. aureus, or S. pneumo-
niae at an MIC concentration of 64 lg/mL, although compounds 2j
and 2k did have moderate activity against the hypersensitive S.
aureus Smith strain (MIC = 4 g/mL and 1.3 g/mL, respectively).
Analysis of the X-ray structures of 1 and additional analogs
above suggested that further exploration of the 5-position of the
Fig. 3. (A) Superposition of X-ray crystal structures of compound 4f (orange) bound
to S. aureus GyrB. (PDE code 3G7B) and compound 4b (black) bound to E. coli GyrB.
(PDE code 3G7E). (B) Portion of the molecular surface of compound 4f shown with
compounds 4f (orange) and 4b (black).

2830
S. M. Ronkin et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2828–2831
Table 2
Ar
Ar
Ar
S
Escherichia coli enzyme activity
c
S
N
N
a
S
N
b
S
N
Ar
NH2
O
O
O
O
OTf
13
OH
11
OTf
12
R
7y Ar = phenyl
O
O
O
7z Ar = 3-pyridyl
N
O
NH
N
S
Ar
Ar
Ar
Ar
Ar
R
E. coli Ki (
lM)
S
N
S
N
S
N
e, f, g
h
d
2a
2b
2c
2d
2e
2f
2g
2h
2i
Phenyl
Phenyl
Phenyl
Phenyl
3-Pyridyl
3-Pyridyl
3-Pyridyl
3-Pyridyl
3-Pyridyl
3-Pyridyl
3-Pyridyl
Propyl
Phenyl
1-Piperidine
4-Morpholine
Methyl
4-Methyl-1-piperidine
Methoxymethyl
Hydroxymethyl
Cyclohexyl
0.8
N
N
O
N
N
HN
N
0.14
0.21
0.63
0.33
1.05
0.28
0.67
0.12
0.24
0.07
HN
N
O
O
O
N
H
O
3b
14
2c
15 Ar = 3-pyridyl
Scheme 2. Reagents and conditions: (a) diethylbromomalonate, pyridine, toluene
reflux; (b) trifluoroacetic anhydride, 2,6-lutidine, CH2Cl2, 0 °C; (c) Me2AlCl, N,O-
dimethylhydroxyl-amine hydrogen chloride, CH2Cl2, 0 °C; (d) piperidine, toluene
reflux; (e) MeMgBr, THF, 0 °C; (f) KOtBu, diethyl oxalate, THF, rt; (g) hydrazine
hydrate, acetic acid, ethanol; (h) ethyl amine, methanol reflux.
2j
2k
Cyclopropyl
Phenyl
explains the 10 times increase in potency in E. coli over S. aureus for
these longer carbamate analogs.
afforded the desired pyrazole esters. The lithium enolate of methyl
ketone 9 was also condensed with propionaldehyde, and subse-
quent oxidation with Dess–Martin periodinane resulted in the de-
sired b-diketone 10a, which was converted to ethylpyrazole 1g.
The nitrogen-linked C-4 thiazole analogs (Scheme 2) were synthe-
sized by first condensing diethylbromomalonate with the aryl-
thioamide 7y or 7z to afford the 4-hydroxythiazole 11 as
described by Kerdesky et al.10 Treatment of 11 with trifluoroacetic
anhydride and 2,6-lutidine afforded triflate 12. The ethyl ester
moiety in 12 was converted to the Weinreb amide 13, and subse-
quent SnAr displacement of the triflate with piperidine afforded
14. The synthesis was completed as described in Scheme 1.
The pyrazole amides 3a–c, 3g described in Table 3 were pre-
pared from the esters 15, 2e, 2i, 2k, respectively, by heating with
ethyl amine in methanol.
The synthesis of the ethyl ester compounds described in Tables
1 and 2 when the 4-position on the thiazole is alkyl or aryl
(Scheme 1), started with the chlorination of the commercially
available b-keto ester 5. In instances when the b-keto ester was
not available from commercial sources, it was prepared via the
route described by Holmquist et al.9 The chlorinated product 6
was condensed with the arylthioamide 7x–z followed by direct
conversion of the ethyl ester to the Weinreb amide 8 using dimeth-
ylaluminum chloride. A Grignard addition afforded the methyl ke-
tone 9, which was subjected to a Claisen condensation to yield the
desired keto ester 10. Subsequent condensation with hydrazine
S
O
O
O
O
b, c
a
Ar NH2
O
R
+
O
R
Cl
5
Table 3
7x Ar = 2-thiophene
7y Ar = phenyl
6
R = aryl or alkyl
Ar
Escherichia coli enzyme activity and antibacterial activity (MIC)
7z Ar = 3-pyridyl
Ar
d
S
N
O
R
S
N
N
O
R
O
N
R
H
N
NH
N
S
O
N
9
8
Ar
Ar
Ar
S
N
g
S
N
R
E. coli Ki (
l
M)
MICa
(lg/mL)
e, f
S
N
O
R
O
N
R
R
3a
3b
3c
Methyl
1-Piperidine
Phenyl
0.35
0.090
0.045
All MIC > 64
HN
dE. coli tol C = 32
bS. aureus = 8
O
9
1g
10a
Ar = 2-thiophene
R = methyl
cS. pneumoniae = 2
dE. coli tol C = 4
H
N
bS. aureus = 32
Ar
N
Ar
N
Ar
N
cS. pneumoniae = 1.5
dE. coli tol C = 48
h
i
3d
0.040
S
S
S
O
O
R
N
R
R
HN
O
H
N
O
O
9
3e
3f
3.4
All MIC >64
10
1f,
2a - 2k
O
Ar
O
S
N
j
H
N
N
R
0.082
dE. coli tol C = 32
HN
dE. coli tol C = 16
bS. aureus = 8
N
H
O
3g
3h
Cyclohexyl
4-Cl-Phenyl
0.062
0.047
3a, 3c
cS. pneumoniae = 4
dE. coli tol C = 8
Scheme 1. Reagents and conditions: (a) SO2Cl2, CH2Cl2, 0 °C to rt; (b) ethanol,
reflux; (c) Me2AlCl, N,O-dimethylhydroxy-amine hydrogen chloride, CH2Cl2, 0 °C;
(d) MeMgBr, THF, 0 °C; (e) lithium bis(trimethylsilyl) amide, propionaldehyde, THF,
À78 °C to À20 °C; (f) Dess–Martin periodinane, t-butyl alcohol, CH2Cl2, 0 °C; (g)
hydrazine hydrate, ethanol; (h) KOtBu, diethyl oxalate, THF, rt; (i) hydrazine, acetic
acid, ethanol; (j) ethyl amine, methanol, reflux.
a
b
c
Observed antibacterial activities.
S. aureus ATCC29213.
S. pneumoniae ATCC10015.
d
E. coli tol C strain CAG12184(CGSC)

S. M. Ronkin et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2828–2831
2831
H2N
Ar
Ar
O
OBn
O
OBn
O
OBn
O
O
N
S
S
N
S
N
Ar
a, b, c
N
Cl
N
d
N
HO
S
N
N
( )n
( )n
N
HN
N
N
HN
N
( )n
a, b
c, d
16a: n = 0
18
17
O
N
16b: n = 1
NH
NH
O
N
S
SnBu3
O
O
O
O
Br
N
R1
O
OBn
N
g
e, f
O
HN
N
+
S
OBn
R1 = methyl 4a, 4f
= 2-butyne 4c-e
= 2 propyne 4b
N
8, 14
21
NH
N
Br
O
Br
( )n
( )n
20
19
Scheme 4. Reagents and conditions: (a) tert-butyl prop-2-ynylcarbamate, n-
butyllithium, THF, À15 °C to À10 °C; (b) hydrazine hydrate, ethanol; (c) TFA,
CH2Cl2; (d) but-2-yn-1-ol (4c–e), 1,1-carbonyldiimidazole, THF.
N
S
N
h, i
H
N
N
potent enzyme and moderate antibacterial activity. They possess
activity against S. aureus and S. pneumoniae, but appear to be ac-
tively effluxed from E. coli based on E.coli tol C mutant activity ver-
sus E. coli wild type activity. Perhaps the most interesting
observation to arise from this study is the ability of the carbamate
analogs to act as selective inhibitors of E. coli GyrB over S. aureus
GyrB. Structural studies explain this selectivity by defining the dif-
ferences of the binding site created by Ile-51, Leu-103, and Ile-175
in S. aureus GyrB compared to E. coli where the isoleucines are va-
lines and Leu-103 is a methionine.
HN
( )n
N
H
O
3d-f
Scheme 3. Reagents and conditions: (a) isobutyl chloroformate, TEA, ethyl acetate,
0 °C to rt; (b) CH2N2/Et2O, ethyl acetate; (c) 6 N HCl, ethyl acetate; (d) thiourea,
ethanol, reflux; (e) NBS, CH3CN, 0 °C; (f) isoamyl nitrite, CuBr2, PEG 200, CH3CN; (g)
3-(1,3,2-dioxborinan-2-yl)pyridine, 2 N aq NaHCO3, Pd(PPh3)4, DME, reflux; (h)
(Ph3P)2PdCl2, dioxane, reflux; (i) BBr3, TFA, CH2Cl2, 0 °C.
The synthesis of the 2-piperidine and 2-pyrrolidine analogs
(Scheme 3) started with commercially available CBZ-protected 2-
carboxylic acid 16, which was activated by making the mixed
anhydride. This was subjected to reaction with diazomethane, fol-
lowed by treatment with 6 N HCl to afford chloromethyl ketone 17.
Condensation of 17 with thiourea afforded the desired 2-aminothi-
azole 18, which was brominated at the 5-position on the thiazole
ring with N-bromosuccinamide. The dibromo compound 19 was
synthesized by aniline diazotization followed by treatment with
copper bromide, using PEG 200 to maintain solubility. Suzuki cou-
pling with 3-(1,3,2-dioxborinan-2-yl)pyridine proceeded selec-
tively at the 2-position of thiazole 20, and subsequent Stille
coupling with the desired pyrazole stannane11 afforded com-
pounds 3d–f.
The synthesis of the carbamate compounds in Table 4 started
with alkynylation of the Weinreb amides (8, 14) (Scheme 4). The
amide was displaced by the dianion of BOC-propargyl amine. Sub-
sequent treatment with hydrazine hydrate afforded pyrazole 21.
The BOC protecting group was cleaved with TFA, and the resulting
amine was converted to the desired carbamates 4a–f, by treatment
with the desired alcohol in the presence of CDI.12
References and notes
1. Barrett, C. T.; Barrett, J. F. Curr. Opin. Biotechnol. 2003, 14, 621.
2. Wolfson, J. S.; Hooper, D. C. Antimicrob. Agents Chemother. 1985, 28, 581.
3. Ho, P-L.; Que, T.-L.; Tsang, D. N.-C.; Ng, T.-K.; Chow, K.-H.; Seto, W.-H.
Antimicrob. Agents Chemother. 1999, 43, 1310.
4. Drlica, K.; Malik, M.; Kerns, R. J.; Zhao, X. Antimicrob. Agents Chemother. 2008,
52, 385.
5. Lewis, I. J.; Singh, O. M. P.; Smith, C. V.; Skarzynski, T.; Maxwell, A.; Wonacott,
A. J.; Wigley, D. B. EMBO J. 1996, 15, 1412.
6. Hull, K.; Green, O.; Sing, A.; Bist, S.; Demeritt, J.; Loch, J.; Mulled, G.; Hauk, S.;
Sherer, B.; Ni, H.; Eakin, A.E. Presented at the 48th Interscience Conference on
Antimicrobial Agents, Washington D.C., October, 2008; Poster F1-2027.
7. Illingworth, R.N.; Uria-Nickelsen, M.; Bryant, J.; Eakin, A.E. Presented at the
48th Interscience Conference on Antimicrobial Agents, Washington D.C.,
October 2008; Poster F1-2028.
8. Charifson, P. S.; Grillot, A.-L.; Grossman, T. H.; Parsons, J. D.; Badia, M.; Bellon,
S.; Deininger, D. D.; Drumm, J. E.; Gross, C. H.; LeTiran, A.; Liao, Y.; Mani, N.;
Nicolau, D. P.; Perola, E.; Ronkin, S.; Shannon, D.; Swenson, L. L.; Tang, Q.;
Tessier, P. R.; Tian, S.-K.; Trudeau, M.; Wang, T.; Wei, Y.; Zhang, H.; Stamos, D. J.
Med. Chem. 2008, 51, 5243.
9. Holmquist, C. R.; Roskamp, E. J. J. Org. Chem. 1989, 54, 3258.
10. Kerdesky, F. A. J.; Holms, J. H.; Moore, J. L.; Bell, R. L.; Dyer, R. D.; Carter, G. W.;
Brooks, D. W J. Med. Chem. 1991, 34, 2158–2165.
11. Sakamoto, T.; Shiga, F.; Uchiyama, D.; Kondo, Y.; Yamanaka, H. Heterocycles
1992, 33, 813–818.
12. For full experimental details on similar compounds see: US Patent 0024030 A1,
2005.
In conclusion, the pyrazolothiazole class evolved from a micro-
molar high throughput screen hit, to a class of GyrB inhibitors with
Table 4
Staphlococcus aureus, Escherichia coli enzyme activity and antibacterial activity (MIC).
O
R
N
S
N
H
OR1
NH
N
Ar
Ar
R
R1
Methyl
CH2C„CH
CH2C„CHCH3
CH2C„CHCH3
CH2C„CHCH3
Methyl
S. aureus Ki(
lM)
E. coli Ki(
lM)
MIC (lg/mL) E. coli tol C
4a
4b
4c
4d
4e
4f
3-Pyridyl
3-Pyridyl
3-Pyridyl
3-Pyridyl
3-Pyridyl
Phenyl
1-Piperidine
1-Piperidine
1-Piperidine
Methoxymethyl
Cyclohexyl
0.300
0.240
0.140
1.10
0.056
0.011
<0.004
0.014
<0.004
1.08
>64
1.0
0.5
8.0
>64
>64
0.069
4-Hydroxy-1-piperidine
NDa
a
Not tested for S. aureus activity