Preparation of novel anthranilic acids as antibacterial agents. Extensive evaluation of alternative amide bioisosteres connecting the A- and the B-rings

Article abstract of DOI:10.1016/j.bmcl.2007.02.055

In the past few years, a significant effort has been devoted by Pharmacia toward the discovery of novel antibiotics. We have recently described the identification of an anthranilic acid lead 1 and the optimization resulting in the advanced lead 2. In this report, we describe the preparation of several selected amide bioisosteres connecting the A- and the B-rings. The E-alkene provided a rigid analog with equal potency to the corresponding amide. This indicates that the amide is not a recognition element rather acts as an appropriate spatial linker of the two important aryl A and B rings. The work here clearly demonstrates that the amide linker can be replaced with several functionalities without significant deterioration in the MIC activity.

Full text of DOI:10.1016/j.bmcl.2007.02.055

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2823–2827
Preparation of novel anthranilic acids as antibacterial agents.
Extensive evaluation of alternative amide bioisosteres
connecting the A- and the B-rings
Atli Thorarensen,*, Brian D. Wakefield, Donna L. Romero, Keith R. Marotti,
Michael T. Sweeney, Gary E. Zurenko, Douglas C. Rohrer, Fusen Han
and Garold L. Bryant, Jr.
Medicinal Chemistry and Infectious Diseases Biology, Pharmacia Corporation, 301 Henrietta Street, Kalamazoo, MI 49001, USA
Received 1 February 2007; revised 19 February 2007; accepted 21 February 2007
Available online 25 February 2007
Abstract—In the past few years, a significant effort has been devoted by Pharmacia toward the discovery of novel antibiotics. We
have recently described the identification of an anthranilic acid lead 1 and the optimization resulting in the advanced lead 2. In this
report, we describe the preparation of several selected amide bioisosteres connecting the A- and the B-rings. The E-alkene provided a
rigid analog with equal potency to the corresponding amide. This indicates that the amide is not a recognition element rather acts as
an appropriate spatial linker of the two important aryl A and B rings. The work here clearly demonstrates that the amide linker can
be replaced with several functionalities without significant deterioration in the MIC activity.
Ó 2007 Elsevier Ltd. All rights reserved.
Emergence of bacterial resistance is a significant prob-
lem in the treatment of bacterial infections,1–3 and this
has fueled a continuous search for novel antibiotics
resulting in numerous commercially available products.
Pharmacia had significant effort devoted to the discov-
ery of novel antibacterial agents over a period of many
years. We recently reported4 on the optimization of acid
1 resulting in the discovery of the lead compound 2
which displayed potent broad-spectrum antibacterial
activity. Subsequent optimization of 2 targeted reducing
its affinity for human serum albumin (HSA) resulting in
advanced compound 3.5
rate hypotheses; (i) that the amide linker had a key inter-
action with the target and disruption would result in
decreased activity or (ii) that the purpose of the amide
linker was to act as a conformational control element
connecting key pharmacophores (Fig. 1).
There are numerous examples where replacing func-
tional groups with alternative bioisosteres was success-
ful.6,7 We selected several functional groups that had
been successfully utilized as amide replacements by
maintaining the key recognition elements of an amide
In the absence of structural information, the relative
importance of each functional group embedded in com-
pounds 2 and 3 was unclear, although prior work
showed that the carboxylic acid was required for activ-
ity. This paper will disclose our work in exploring the
role of the linker connecting the A- and the B-rings.
The design of amide isosteres was guided by two sepa-
O
CO2H
Br
CO2H
A
B
NC
OH
NH
NH
O
NH
O
N
O
O
O2S
N
O2S
N
2
BnS
3
Keywords: Antibacterial; Bioisoster; Amide; Cyclopropane.
*
1
Corresponding author. Tel.: +1 636 247 3962; fax: +1 636 247
0250; e-mail: atli.thorarensen@pfizer.com
Cl
Figure 1. Progression in lead optimization.
Present address: Pfizer, 700 Chesterfield Parkway West AA2G,
Chesterfield, MO 63017, USA.
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.02.055

2824
A. Thorarensen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2823–2827
either by incorporating hydrogen accepting or donating
properties. The design of novel congeners was aided by
extensive modeling overlaying the potential linking moi-
eties. The modeled subunits were overlaid with amide 4
and a similarity score was calculated. The modeling indi-
cated that the linkers that were planned had a conforma-
tion available very similar to the ground state
conformation of 4, Table 1. A priori the importance of
the differences in the similarity scores was hard to judge.
Therefore, a wide range of new linkers were selected and
prepared.
depending on the ease of synthesis, since the initial
objective was to illustrate the ability of the linker to re-
place the amide. It should be recognized that the selected
substitution would not illustrate the full potential of the
corresponding linker.
The thioamide was readily available from the methyl
ester 8 and Lawsson’s reagent in refluxing toluene,
Scheme 1.12 The majority of the Lawsson’s reagent
could be removed by filtration; however, several silica
gel flash chromatographies were required to completely
purify the product. The ester was hydrolyzed under
standard conditions to afford the orange thioamide 9.
Our goal from the outset was the systematic preparation
of analogs containing a linking unit using comparable
fragments on the A- and the B-rings. It should be noted
that prior work showed that electron withdrawing
groups at the 5-position on the A ring such as –Cl,
–Br, –NO2, and –CN were interchangeable and had sim-
ilar antibacterial activity.4,11 Likewise, prior work had
also indicated that 3- versus 4-substituted B-rings pro-
vide analogs with similar activity, with the 3-substitu-
tion providing slightly improved potency. In addition
the selection of a B-ring substituent was guided by the
fact that the corresponding amide analog had illustrated
reasonable antibacterial activity. For the purposes of
this work, the substitution patterns were selected
O
O
1. Lawson's Rgnt,
tol (reflux)
NC
NC
OH
NH
O
2. LiOH, THF, H2O
NH
55%
SO2R
SO2R
S
O
N
R =
8
9
Cl
Scheme 1.
Table 1. Selected examples of the overlays utilized in design and selection of linkers
Compound
Structure
#Conf similarity
Best similarity
O
Br
OH
4
Model
NH
O
O
Br
OH
NH
5
12
0.69
S
O
Br
OH
6
11
0.66
O
Br
OH
7
7
0.63
The minimum energy conformation closest to that of the model compound 2 was obtained through conformational searching retaining those within
3 kcal/mol of the lowest energy8 followed by ab initio QM geometry optimization at the HF/3-61G level.9 The resulting analog structures were
overlaid upon the structure of compound 2 and the pair-wise similarities were evaluated using a field-based similarity technique.10

A. Thorarensen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2823–2827
2825
Removing the linking group between the A- and the
B-ring results in contracted analogs where the carboxyl-
ate could be positioned on either ring. The analog
containing the A-ring carboxylate (14) was prepared in
excellent overall yield via a Stille coupling followed by
hydrolysis (Scheme 2). Incorporation of the carboxylate
on the B-ring was performed in a similar manner,
Scheme 3. The installment of the carboxylate on the
B-ring and removing the amide allows the location of
the acidic proton in a very similar space while retaining
other features. In this instance, the organometallic part-
ner is boronic acid 16 which was coupled in a Suzuki
reaction affording the core subunit 17 in 80% yield.
Hydrogenation of the nitro group afforded amine 18,
treatment with tolylsulfonylchloride provided 20 which
was deprotected to afford the final analog 21 in excellent
overall yield.
24 was easily prepared in several steps, using sequential
Sonogashira couplings. Unfortunately hydrolysis and
selective partial hydrogenation were unsuccessful, how-
ever elevated pressure (50 psi) and prolonged reaction
time (24 h) afforded the fully saturated ethylene linked
analog 25.
Due to the difficulty with the selective hydrogenation, an
alternative route was designed to provide the E-alkene
linker, Scheme 5. This approach used phosphonylide
chemistry which we had previously enjoyed success with
in efforts to prepare efflux pump inhibitors.14 First
the requisite nitrile was prepared by hydrogenation of
the corresponding nitro compound 26, followed by
I
Cl
There are a number of amide bioisosteres described in
the literature and among them E-alkenes have been
successfully utilized to replace amides.13 Our approach
accessing a C–C linked A- and B-ring was based on
sequential Sonogashira couplings, Scheme 4. This ap-
proach has the potential of providing a series of analogs
with differing degrees of saturation via sequential selec-
tive hydrogenations. The fully assembled alkynyl analog
O
Cl
1) H+ / MeOH
SO2
O
N
Cl
88%
TMS
O
2)
O
10
CO2H Pd(PPh3)2Cl2
100 %
Cl
23
I
SO2
3) KF/ K2CO3 / MeOH
22
N
Pd(PPh3)2Cl2
90%
68%
24
1) PtO2
HO
O
98%
X
O2
S
N
Cl
Cl
2) NaOH
89%
Cl
Cl
Cl
OMe
Cl
25
OR
O2S
N
Br
12
O
Scheme 4.
O
Pd(PPh3 4
)
Pd(PPh3)4
83%
X = I
10
X = SnMe3
(Me3Sn)2
O
Cl
1) H+ / MeOH
93%
11
O
O2S
1) NBS / AIBN
OMe
NC
N
63%
O2N
76%
OH
2) Raney Nickel H2
100%
3) NaNO2 / HCl
13
14
R = Me
NaOH
42%
2) PPh3 / toluene
82%
27
26
O
R = OH
CuCN / KCN
NC
42 %
OR
Scheme 2.
O
NaH
NC
OMe
73%
1/1 E/Z
Cl
Cl
Cl
28
PPh3Br
Cl
O
O2S
Br
O
N
16
O
SO2Cl
19
OMe
R = OMe
B(OH)2
OMe
1) NaOH
95%
2) separation
N
30
S
pyridine
64%
Pd(PPh3)4
80%
R = OH
trans
31
Cis
32
O2
29
15
NO2
X
O
NC
Raney Nickel
100%
1) Pddba CH2N2
95 %conversion
of trans only
OH
X = NO2
17
X = NH2
18
20
R = Me
2) O3
3) NaOH
6%
NaOH
95%
HN
SO2
Cl
33
Cl
21 R = OH
O2S
O
N
RO
Scheme 3.
Scheme 5.

2826
A. Thorarensen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2823–2827
diazotization and cyanation. Bromination and reaction
with triphenyl phosphine provided phosphine 28.
Deprotonation with NaH and condensation with the
fully functionalized aldehyde 29 afforded the desired
alkene 30 as a 1:1 mixture of E/Z isomers. The alkene
30 was hydrolyzed and separated to provide the pure
E (31) and Z (32) isomers.
prepared in a similar but lower yielding sequence (not
shown but in vitro activity was similar to that of the
truncated analog 39, validating our approach that only
linkers that illustrated the potential to replace the amide
justified the preparation of the more complex analog for
direct comparison across multiple linkers). Unfortu-
nately, the fully functionalized oxadiazole analog
resisted any oxidation.
Since a trans substituted cyclopropane yields a very sim-
ilar spatial arrangement to the amide and the alkene, al-
kene 30 was converted to a cyclopropane. It was
discovered after extensive attempts that alkene 30 was
very unreactive. Screening of over 30 methylene transfer
conditions found that only Pddba with excess diazo-
methane was able to achieve cyclopropanation with
complete conversion of the desired E-alkene.15 The
Z-alkene was found to be inert to cyclopropanation
under any conditions. Attempts to separate the remaining
Z-alkene failed, thus ozonization was utilized to convert
the remaining alkene to readily separated materials. The
desired trans cyclopropane 33 was obtained in poor
yields due to decomposition during the ozonolysis but
with high purity.
The ability of the new linker as amide replacement was
monitored by its activity against Staphylococcus aureus.
Not surprisingly we found that the thioamide 9 is equi-
potent with the corresponding amide, Table 2. It is
intriguing that the spectrum of activity for the thioamide
is improved though the reason for that activity improve-
ment is not clear. The conversion of an amide to a thio-
amide represents the smallest possible change in linker
construction. The removal of the linker as in 14 and
21 results in analogs devoid of activity. Likewise, when
the linker is replaced with five-membered ring heterocy-
cles such as 37 and 39 all activity is lost. A flexible linker
such as alkane 25 exhibits faint antibacterial activity.
The alkenes provide a rigid scaffold and the activity of
the E-alkene 31 is equal to the potency of the corre-
sponding amide. This indicates that the amide is not
essential for activity and serves to hold the pharmaco-
phores at the appropriate distances. Therefore, it is
not surprising that the corresponding cyclopropyl linker
found in 33 is nearly as active as the corresponding
E-alkene 31. The Z-alkene 32 possessed poor activity
indicating that the A- and the B-rings were inappropri-
ately positioned.
Another linker that has been reported in the literature as
a viable bioisostere for amides is a triazole.16 Modeling
indicates that the triazole can occupy similar conforma-
tions as the amide, although the B-ring would be slightly
shifted. In addition the triazole retains the hydrogen
bonding pattern of the amide. In order to prepare this
linker, we initiated a model study for its assembly,
Scheme 6. Fortunately, classical literature observations
were reproduced and depending on whether the hydro-
chloride salt of benzimidate was used or not we obtained
either the triazole or oxadiazole as single isomers out of
each reaction in modest yields.17 Conversion of the
methyl groups to the acids was troublesome and was
achieved with chromium oxidation for the oxadiazole
36, and under basic conditions with KMnO4 for triazole
38.18 The fully functionalized triazole analog was
The work described herein clearly demonstrates that the
purpose of the amide linker is to appropriately position
the aryl rings and it does not have a role as a recognition
element. The amide can be replaced with several
functionalities such as the E-alkene and the trans cyclo-
propyl without significant deterioration in the MIC
activity. These findings have enabled us to propose a
NH
Table 2. Antibacterial activity of new linkers in this report
HCl
Br
R
Compound
MICa (lg/mL)
Ph
OMe
35
SAURb
0.5
EFAEc SEPIdd
SPNEe HINFf
O
Xylene
reflux
60%
2
9
32
32
4
0.125
4
16
128
64
N
Br
N
0.125
36
R = Me
1) (COCl)2
14
21
25
31
32g
33
37
39
128
>128
32
>128
>128
>128
128
128
>128
>128
32
>128
>128
>128
>128
>128
16
NaCr2O7
10%
64
32
2) NH2NH2
20%
R = CO H
0.25
0.5
>128
64
O
OH
37
2
NH
HCl
34
16
4
>128
>128
>128
>128
32
8
>128
>128
>128
Ph
OMe
Br
R
35
>128
>128
>128
>128
>128
>128
H
N
Dowex-OH
Xylene
reflux
a Minimal inhibitory concentration.
b Staphylococcus aureus UC 9218.
c Enterococcus faecalis UC 9217.
d Staphylococcus epidermidis UC 12084.
e Streptococcus pneumoniae UC 9912.
f Haemophilus influenzae 30063.
N
N
38 R = Me
74%
KMnO4 / NaOH
38%
39 R = CO2H
g This is a 97.9/2.1 cis/trans mixture and the activity can be mainly
attributed to the trans impurity.
Scheme 6.

A. Thorarensen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2823–2827
2827
9. Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S.
T.; Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga,
N.; Nguyen, K. A.; Su, S. J.; Windus, T. L.; Dupuis, M.;
Montgomery, J. A. J. Comput. Chem. 1993, 14, 1347.
10. Mestres, J.; Rohrer, D. C.; Maggiora, G. M. J. Mol.
Graphics Model. 1997, 15, 114.
pharmacophore model for this series of compounds, de-
tails of which will be the topic of future publication.
References and notes
11. (a) Ruble, J. C.; Thorarensen, A, unpublished results; (b)
The antibacterial activity for numerous substitution pat-
terns is provided in WO2004018414.
12. Cava, M. P.; Levinson, M. I. Tetrahedron 1985, 41,
5061.
13. Tamamura, H.; Yamashita, M.; Nakajima, Y.; Sakano,
K.; Otaka, A.; Ohno, H.; Ibuka, T.; Fujii, N. J. Chem.
Soc. Perkin Trans. 1999, 1, 2983.
14. (a) Thorarensen, A.; Presley-Bodnar, A. L.; Marotti, K.
R.; Boyle, T. P.; Heckaman, C. L.; Bohanon, M. J.;
Tomich, P. K.; Zurenko, G. E.; Sweeney, M. T.; Yagi, B.
H. Bioorg. Med. Chem. Lett. 2001, 11, 1903; (b) Vedejs, E.;
Fleck, T. J. J. Am. Chem. Soc. 1989, 111, 5861.
15. Denmark, S. E.; Stavenger, R. A.; Faucher, A.-M.;
Edwards, J. P. J. Org. Chem. 1997, 62, 3375.
1. Cohen, M. L. Science 1992, 257, 1050.
2. (a) Spratt, B. G. Science 1994, 264, 388; (b) Nikaido, H.
Science 1994, 264, 382.
3. Cunha, B. A. Drugs Today 1998, 34, 691.
4. Larsen, S. D.; Hester, M. R.; Ruble, J. C.; Kamilar, G.
M.; Romero, D. L.; Wakefield, B.; Melchior, E. P.;
Sweeney, M. T.; Marotti, K. R. Bioorg. Med. Chem. Lett.
2006, 6173.
5. Li, J.; Wakefield, B. D.; Ruble, J. C.; Stiff, C. M.; Romero,
D. L.; Marotti, K. R.; Sweeney, M. T.; Zurenko, G. E.;
Rohrer, D. C.; Thorarensen, A. Bioorg. Med. Chem. Lett.
2007, ASAP.
6. Patani, G. A.; LaVoie, E. J. Chem. Rev. 1996, 96, 3147.
7. Chen, X.; Wang, W. Annu. Rep. Med. Chem. 2003, 38, 333.
8. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.;
Liskamp, R.; Lipton, M.; Caufield, C.; Chang, G.;
Hendrickson, T.; Still, W. C. J. Comput. Chem. 1990, 11,
440.
16. Sui, Z. et al. Bioorg. Med. Chem. Lett. 1998, 8, 1929.
17. Weidinger, H.; Kranz, J. Chem. Ber. 1963, 96, 1049.
18. Harris, R. L. N.; Huppatz, J. L. Aust. J. Chem. 1977, 30,
2225.