The synthesis and biological evaluation of novel series of nitrile-containing fluoroquinolones as antibacterial agents

Article abstract of DOI:10.1016/j.bmcl.2007.01.090

Several novel series of nitrile-containing fluoroquinolones with weakly basic amines are reported which have reduced potential for hERG (human ether-a-go-go gene) channel inhibition as measured by the dofetilide assay. The new fluoroquinolones are potent

Full text of DOI:10.1016/j.bmcl.2007.01.090

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2150–2155
The synthesis and biological evaluation of novel series
of nitrile-containing fluoroquinolones as antibacterial agents
Sean T. Murphy,a,* Heather L. Case,b Edmund Ellsworth,b Susan Hagen,b
Michael Huband,b Themis Joannides,b Chris Limberakis,b Keith R. Marotti,b
Amy M. Ottolini,b Mark Rauckhorst,b Jeremy Starr,b Michael Stier,b
Clarke Taylor,b Tong Zhu,b Adrian Blaser,c William A. Denny,c
Guo-Liang Lu,c Jeff B. Smaillc and Freddy Rivaultd
aPfizer Global Research and Development, La Jolla Laboratories, 10770 Science Center Drive, San Diego, CA 92121, USA
bPfizer Global Research and Development, Ann Arbor Laboratories, 2800 Plymouth Road, Ann Arbor, MI 48105, USA
cAuckland Cancer Society Research Centre, School of Medical Sciences, The University of Auckland,
Private Bag 92019, Auckland 1020, New Zealand
dMetaux et Microorganismes: Chimie, Biologie et Applications, Departement des Recepteurs et Proteines Membranaires,
UMR7175 Institut Gilbert Laustriat, Ecole Superieure de Biotechnologie de Strasbourg,
Boulevard Sebastien Brant, F-67400 Illkirch, France
Received 13 October 2006; revised 24 January 2007; accepted 29 January 2007
Available online 2 February 2007
Abstract—Several novel series of nitrile-containing fluoroquinolones with weakly basic amines are reported which have reduced
potential for hERG (human ether-a-go-go gene) channel inhibition as measured by the dofetilide assay. The new fluoroquinolones
are potent against both Gram-positive and fastidious Gram-negative strains, including Methicillin resistant Staphylococcus aureus
and fluoroquinolone-resistant Streptococcus pneumoniae. Several analogs also showed low potential for human genotoxicity as mea-
sured by the clonogenicity test. Compounds 22 and 37 (designated PF-00951966 and PF-02298732, respectively), which had good
in vitro activity and in vitro safety profiles, also showed good pharmacokinetic properties in rats.
Ó 2007 Elsevier Ltd. All rights reserved.
Antibacterial resistance is a growing problem1,2 which
necessitates the discovery of new antibiotics with activity
against resistant strains. The fluoroquinolone class of
antibiotics is an important weapon against bacterial
infections, in particular Ciprofloxacin and Levofloxacin
(see Fig. 1), but resistance is increasing in this class as
well.3 Some more recently approved fluoroquinolones
have had concerns with cardiovascular safety, in partic-
ular the possibility of QTc prolongation which can lead
to potentially fatal torsades de pointes.4,5 These concerns
led to Sparfloxacin and Grepafloxacin being removed
from the market and bolded warnings being added to
the package inserts for Moxifloxacin and Gatifloxacin.
Due to the serious consequences of inducing QTc pro-
longation, much work has been done to understand
the effect and to develop various methods to predict
it.6 Typical in vitro assays include a hERG (human
ether-a-go-go related gene) functional assay and a dof-
etilide binding assay which measures the ability of a
compound to displace dofetilide—a potent inhibitor in
the functional assay—from the hERG channel. The dof-
etilide assay has a higher throughput than the hERG
functional assay and is thus more useful at the discovery
stage.6 Several pharmacophore models have been devel-
oped for hERG activity, and the general findings indi-
cate that a basic nitrogen and two or three aromatic
components at appropriate distances from the amine
confer increased hERG activity.6 We set out to decrease
the hERG activity by decreasing the basicity of the
amine with the electron-withdrawing nitrile group, while
at the same time maintaining activity against resistant
organisms.
Keywords: Fluoroquinolones; Antibacterial; Resistance; Nitriles; Dof-
etilide; hERG; pKa; PF-00951966; PF- 02298732; PF-02789402.
*
Corresponding author. E-mail: sean.murphy2@pfizer.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.01.090

S. T. Murphy et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2150–2155
2151
O
N
O
N
O
N
O
O
NH2
O
F
F
F
OH
OH
OH
OH
N
N
N
HN
N
O
HN
F
Ciprofloxacin
Levofloxacin
Sparfloxacin
O
O
O
N
O
O
O
F
F
F
OH
OH
H
N
N
N
N
N
N
HN
HN
O
O
Grepafloxacin
Gatifloxacin
Moxifloxacin
Figure 1. Chemical structures of some approved fluoroquinolones.
Our initial series was derived from the C7-pyrrolidine
side chains (see Table 1 for fluoroquinolone numbering)
developed by Domagala et al.7 The previous series had
excellent activity against resistant strains but showed
significant displacement of dofetilide in the dofetilide
binding assay. The pyrrolidine side chain series also
showed various levels of mammalian genotoxicity, so
we also monitored the compounds in a clonogenicity as-
say which has been shown to be predictive in the fluor-
oquinolone class.8
amine surrogate (azide, phthalimide, carbamic acid
tert-butyl ester) to 4 by a Michael addition were unsuc-
cessful, we found that either ammonia or methyl amine
could be added directly to 4 with excess amine (5–10
equiv) in ethanol at 80 °C in a sealed tube. After Boc
protection of the amine, the diastereomers were separat-
ed by silica gel chromatography, and then the enantio-
mers of the slower eluting diastereomer were separated
by preparative chiral HPLC to give the desired stereo-
isomer (vide infra).11 After Cbz-deprotection of the pyr-
rolidine, the side chain was coupled to various
fluoroquinolone cores to give 20, 22, 27, 32, and 33 (Ta-
ble 1).
We have found that incorporating a nitrile group either
onto the carbon framework of the pyrrolidine side chain
(C-substituted series, R = CH2CN in Table 1, e.g., 20) or
as a substituent on the nitrogen (N-substituted series,
R0 = CH2CH2CN, e.g., 17) significantly increased the
Ki in the dofetilide assay. Further exploration of ni-
trile-containing fluoroquinolones led to the discovery
of compounds with either pyrrolidine or azetidine side
chains which have activity against key Gram-positive
and Gram-negative bacteria along with improved in vi-
tro safety measures.
The absolute stereochemistry of the desired stereoisomer
of
5 ((S)-3-amino-3-(R)-pyrrolidin-3-yl-propionitrile)
was confirmed to be the same as in the des-nitrile case7
by an independent, stereoselective synthesis (Scheme 3).
The aldehyde 7, synthesized by the method of Baldwin
et al.12 was converted to amine 8 by reductive amina-
tion. Removal of the TBS protecting group from 8 fol-
lowed by heating at reflux in THF induced
a
translactamization from the b-lactam to the c-lactam,
which formed the pyrrolidine ring with the desired ste-
reochemistry. After global reduction and Boc protection
of the amine, the N-benzyl group was replaced with a
Cbz protecting group, the swap being necessary for later
removal in the presence of a nitrile. The nitrile was then
introduced via the mesylate to give side chain 5a, the ste-
reochemistry of which was confirmed to be that of the
desired stereoisomer by chiral HPLC. Methylation of
5a led to 5b which was also confirmed by chiral HPLC
to be the desired stereoisomer in the N-methyl analogs.
The fluoroquinolones with pyrrolidine side chains were
prepared as described previously for the non-nitrile-con-
taining analogs,7,9 an example of which is shown in
Scheme 1. The 8-Me fluoroquinolone analog (Table 1,
28) could be prepared in higher yield by using DMSO
instead of acetonitrile and Hunig’s base rather than tri-
ethylamine. In the case of the analogs containing azeti-
dine side chains (34 and 36), the final deprotection
(removal of trifluoro-acetate, see Scheme 4) was accom-
plished with sodium carbonate in methanol/water at
70 °C. To prepare the analogs where the nitrogen was
substituted with the ethylcyano group (Table 1,
R0 = CH2CH2CN, e.g., 17), the unsubstituted fluoroqu-
inolone (e.g., 16) was heated with excess acrylonitrile
(ca. 5 equiv) and triethylamine in methanol at 50 °C.
The azetidine side chains were synthesized as shown in
Scheme 4 following the route developed by Frigola
et al.13 starting from the known tertiary alcohols 12a13
(R = Me) and 12b14 (R = Et). After reaction with meth-
anesulfonyl chloride, the resulting mesylate of the alco-
hol 12 could be efficiently transformed into the nitrile
with sodium cyanide, provided the solvent was a mix-
ture of DMF and water—running the displacement in
anhydrous DMF gave inconsistent and generally poor
yields. The nitrile 13 was then reduced to the amine,
which was protected as the trifluoroacetate. Finally,
hydrogenolysis of the benzhydryl group freed the azeti-
The synthesis of the nitrile-containing pyrrolidine side
chains 6a and 6b is shown in Scheme 2. The ester 2, pre-
pared by the method of Yoshiyasu et al.10 was converted
by standard procedures into the Cbz-protected aldehyde
3, which was then transformed into the nitrile olefin 4 as
a mixture of isomers by a Horner–Emmons–Wadsworth
olefination. After our initial attempts to introduce an

Table 1. Structure–activity relationships in nitrile-containing fluoroquinolones
O
O
5
4
R6
OH
6
H
N
3
2
R'
7
N
N 1
R8 R1
8
R
H
Compound
R
R0
R6
R8
R1a
Staphylococcus
aureus MICb
(lg/mL)
Staphylococcus Streptococcus Streptococcus Haemophilus Moraxella Dofet.h
Clonoi IC50
aureusRMICc
(lg/mL)
pneumoniae
MICd
pneumoniaeR influenzae
catarrhalis KI (lM) (lg/mL)
MICe
MICf
MICg
(lg/mL)
(lg/mL)
(lg/mL)
(lg/mL)
Ciprofloxacin
Levofloxacin
0.125
0.125
0.015
0.03
1
0.5
0.5
0.5
16
16
0.008
0.015
0.015
0.125
0.004
0.06
0.03
0.06
0.125
0.5
>100 500
>150 (10) >500
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
H
H
CH2CH2CN
F
F
F
F
F
F
F
H
H
H
H
H
F
F
F
F
F
F
OMe cp
OMe cp
OMe cp
OMe cp
OMe cp
OMe cp
OMe cp
OMe cp
OMe cp
OMe cp
OMe cp
OMe cp
0.06
0.06
0.008
0.008
0.002
0.06
0.06
0.25
0.25
0.03
0.125
0.06
0.03
0.008
0.125
0.03
0.06
0.125
0.06
1
24
106
56
252
30
H
0.125
0.004
0.015
0.008
0.008
0.06
2
Me
Me
CH2CN
Me
H
CH2CH2CN
0.002
0.004
0.001
0.015
0.008
0.03
0.03
0.125
0.125
0.125
0.5
4
0.015
0.06
0.03
0.06
0.06
0.5
85
>150 (2)
>150 (29)
>150 (40) >125
48
H
Me
0.008
0.015
0.03
24
CH2CN Me
H
>150 (9)
64
>100
139
331
326
314
38
H
CH2CH2CN
0.125
0.125
0.015
0.25
0.06
0.25
H
Me
0.03
0.008
0.03
2
1
H
CH2CH2CN
0.125
2
0.125
0.5
0.06
0.06
1
62
>100
Me
CH2CN
Me
Me
H
0.008
0.002
0.002
0.015
0.015
0.008
0.06
0.03
0.5
0.03
0.06
0.25
0.25
0.125
2
0.125
0.015
0.015
0.06
0.125
0.125
0.25
>150 (30)
51
72
465
22
H
CH2CH2CN
Me
Me
cp
cp
0.002
0.008
0.03
0.004
0.03
27
Me
Me
H
OCH2CHMe
OCH2CHMe
OCH2CHMe
OCH2CHMe
0.008
0.06
>100 (35)
>150 (10)
>150 (22)
>150 (19)
143
210
159
279
CH2CH2CN
H
0.06
0.015
0.125
CH2CN
CH2CN Me
0.008
0.03
O
O
R6
OH
R'
NH
N
N
R8 R1
R
34
35
36
37
Me
Me
Et
H
F
F
F
F
Me
Me
Me
Me
cp
cp
cp
cp
0.06
0.03
0.015
0.03
0.06
0.06
2
2
1
2
0.015
0.06
0.06
112
>150 (18)
45
112
268
50
CH2CH2CN
H
CH2CH2CN
0.03
0.03
0.03
0.125
0.06
0.125
0.125
0.125
0.25
0.015
0.06
Et
>150 (25)
253
a cp = cyclopropyl, OCH2CHMe between R1 and R8 is the (S)-stereoisomer as in Levofloxacin.
b Staphylococcus aureus UC-76.
c Methicillin resistant Staphylococcus aureus SA-2552.
d Streptococcus pneumoniae SV-1.
e Fluoroquinolone resistant Streptococcus pneumoniae SP-3765.
f Haemophilus influenzae HI-3542.
g Moraxella catarrhalis BC-3531.
h See Ref. 17 for assay conditions, number in parentheses is % inhibition at 300 lM.
i Clonogenicity, see Ref. 8 for assay conditions.

S. T. Murphy et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2150–2155
2153
dine nitrogen for subsequent coupling to the 8-Me fluor-
oquinolone core to give 34 and 36 (Table 1).
O
N
O
O
N
O
F
F
F
OBF2
a-c
OH
Boc
H2N
NC
+
NH
HN
N
NC
H
The new fluoroquinolones were tested for activity against
key pathogens for both community and hospital
infections (Table 1), including Methicillin-resistant
Staphylococcus aureus, Ciprofloxacin-resistant Strepto-
coccus pneumoniae, and the fastidious Gram-negative
species Haemophilus influenzae and Moraxella catarrhalis.15
In the SAR portion of this paper, we will focus on the
resistant S. pneumoniae species since the MIC values for
the other species were generally less than or equal to our
target value of 0.5 lg/mL.
O
O
H
6a
1
20
Scheme 1. Example coupling of side chain 6a with fluoroquinolone
core 1; Reagents and conditions: (a) triethylamine, acetonitrile, 80 °C;
(b) triethylamine, ethanol, 80 °C (79%, 2 steps); (c) HCl, dioxane,
CH2Cl2 (84%).
O
O
NC
a, b, c
d
EtO
H
N
N
N
Cbz
Bn
4
Cbz
The SAR of various pyrrolidine side chains has been
previously discussed by Domagala et al.7 The addition
of a nitrile in the C-substituted series decreased the
activity in S. pneumoniaeR by 4-fold (20 vs 18, 22 vs
21, and 27 vs 25) except in the case of the levofloxacin
core where it showed no decrease in activity (32 vs 30).
The N-substituted pyrrolidine series generally showed
a 2-fold decrease in activity for the R6 = F cores (e.g.,
16 vs 17, 28 vs 29), whereas the R6 des-fluoro core
showed a 16-fold decrease in activity with R = Me (25
vs 26) but no significant change with R = H (23 vs 24).
3
2
e, f
Boc
R
Boc
R
N
N
N
N
g
N
N
6a: R = H
5a: R = H
H
Cbz
6b: R = Me
5b: R = Me
Scheme 2. Reagents and conditions: (a) CbzCl, CHCl3 (85%); (b)
NaBH4 (56%); (c) TEMPO, NaOCl (78%); (d) (EtO)2POCH2CN,
Cs2CO3 (74%); (e) RNH2, EtOH, sealed tube, 80 °C; (f) Boc2O (83–
85%, 2 steps); (g) Pd/C, H2, THF (quant).
The SAR of the azetidine side chains has been examined
by Frigola et al.13 in which the side chain of 34 was test-
ed on the R8 = H and R8 = F cores. In that study, the
addition of the R = Me group versus R = H led to
equivalent activity with R8 = H and about a 4-fold de-
crease in activity with R8 = F. In our series, we were able
to increase the activity by switching to the R8 = Me core
and using R = Et on the side chain to give sufficient
activity against the S. pneumoniaeR strain (1 lg/mL).16
As with the N-substituted pyrrolidine series, the
N-substituted azetidine compounds also showed only a
small decrease in activity compared to the unsubstituted
analogs (34 vs 35, 36 vs 37).
Bn
HN
O
NH2
O
CO2t-Bu
TBS
t-BuO2C
a
CO2t-Bu
TBS
b
H
N Bn
N
N
O
9
O
7
8
c, d, e, f
Me
NC
Boc
H
Boc
Boc
HN
OH
N
HN
H
i
g,h
N Cbz
H
N Cbz
N Cbz
NC
5b
5a
10
In every case where a nitrile was added to the side chain,
the new compound showed less displacement of dofeti-
lide in the dofetilide binding assay17 compared to the
des-nitrile case (e.g., 16 vs 17, 18 vs 20). For the
R8 = Me core, the increase in Ki is only modest (28 vs
29). Since the R8 = Me core is the most lipophilic core
(analog 29 clogD @7.0 is 0.7 higher than for the
R8 = OMe core analog 19), the greater affinity for the
hERG channel18 may be influenced more by lipophilic-
ity than amine basicity.19 In some cases, the des-nitrile
Scheme 3. Reagents and conditions: (a) BnNH2, AcOH; then
NaBH3CN; (b) NaF, MeOH; then reflux in THF/MeOH, (40%, 2
steps); (c) LiAlH4, (d) Boc2O (60%, 2 steps); (e) Pd/C, H2; (f) CbzCl
(50%, 2 steps); (g) MsCl; (h) NaCN (95%, 2 steps); (i) NaH, DMF;
then MeI (72%).
Ph
N
Ph
N
Ph
Ph
a, b
HO
NC
13a: R=Me
13b: R=Et
R
R
and
nitrile-containing
analogs
both
showed
12a: R=Me
c, d
Ki > 150 lM, and therefore the %inhibition at 300 lM
was used for comparison (21 vs 22, 30 vs 31–33). Three
of the new fluoroquinolones (19, 22, and 31) showed
inhibition in the dofetilide assay at 300 lM as low as
Levofloxacin (10%), a marketed fluoroquinolone with
low risk for QTc prolongation.4 Two of these (22 and
31, designated PF-00951966 and PF-02789402) also
met the criteria for clonogenicity—vida infra.
12b: R=Et
Ph
Ph
NH
e
N
CF3CO N
R
CF3CO N
R
H
H
14a: R=Me
14b: R=Et
15a: R=Me
15b: R=Et
Scheme 4. Reagents and conditions: (a) MsCl, triethylamine (R = Me:
96%, R = Et: 98%); (b) NaCN, 4:1 DMF/H2O, 60 °C (R = Me: 75%,
R = Et: 86%); (c) LiAlH4, THF, reflux (R = Me: 95%, R = Et: 92%);
(d) (CF3CO)2O (R = Me: 98%, R = Et: 53%); (e) Pd/C, H2, MeOH,
HCl.
The increase of the Ki in the dofetilide assay was predicted
from the reduced basicity of the amines in the nitrile-con-
tainingcompounds. The reduced basicity was measured 20

2154
S. T. Murphy et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2150–2155
Table 2. Rat PK on selected nitrile-containing fluoroquinolonesa
Compound
AUC, iv (lg h/mL)
CL, iv (mL/min/kg)
Vd, iv (L/kg)
t1/2, iv (h)
Cmax, po (lg/mL)
%F, po
19
22
27
34
37
3.5
4.4
5.2
2.5
6.5
24
19
16
33
13
2.5
1.8
0.5
2.5
1.0
2.7
2.4
1.8
1.3
1.4
1.2
1.4
1.4
0.3
2.6
70
53
22
18
92
a Rats were dosed at 5 mg/kg orally (po) or 1 mg/kg intravenously (iv). All values are an average of at least n = 3.
in the C-substituted series (pKa = 9.5 vs 6.7 for 21 vs 22),
in the N-substituted pyrrolidine series (pKa = 9.9 vs 6.9
for 18 vs 19), and in the N-substituted azetidine series
(pKa = 8.6 vs 6.5 for 34 vs 35).
By modifying the fluoroquinolone core and the side
chain nitrogen substituents, we were able to attenuate
the genotoxicity and discover several compounds which
had both potency and good safety profiles in our in vitro
assays. Further work on the activity of these compounds
in an expanded panel of organisms and in vivo efficacy
models will be reported in due course.
Clonogenicity has been shown to be a useful predictor of
human genotoxicity in the fluoroquinolone class.21 Suto
et al.8 have shown that the clonogenicity in the fluoroqu-
inolone class is affected by the nature of the core as well
as the substituent R0 on the amine (see Table 1). We
sought to control the clonogenicity using similar chang-
es to obtain a compound with an IC50 of >100 lg/mL.
Acknowledgments
We thank Ziqiang Wang for the preparative chiral
HPLC separation of the nitrile side chains.
Compared to the R8 = OMe core, the R8 = Me core ana-
logs were more genotoxic in the clonogenicity assay (19
vs 29), as was observed in the previous study.8 In
contrast, changing to the R6 = H (des-F) core or the
R8–R1 = (S)-OCH2CHMe (Levofloxacin) core generally
improved the clonogenicity (e.g., 20 vs 27 and 32), and
in many cases the target criterion of >100 lg/mL was
satisfied.
References and notes
1. Talbot, G. H.; Bradley, J.; Edwards, J. E., Jr.; Gilbert, D.;
Scheld, M.; Bartlett, J. G. Clin. Infect. Dis. 2006, 42, 657.
2. Theuretzbacher, U.; Toney, J. H. Curr. Opin. Investig.
Drugs 2006, 7, 158.
3. Bakkan, J. S. Scand. J. Infect. Dis. 2004, 36, 85.
4. Kang, J.; Wang, L.; Chen, X.-L.; Triggle, D. J.; Rampe,
D. Mol. Pharmacol. 2001, 59, 122.
5. Owens, R. C., Jr. Drugs 2004, 64, 1091.
6. Jalaie, M.; Holsworth, D. D. Mini-Rev. Med. Chem. 2005,
5, 1083.
7. Domagala, J. M.; Hagen, S. E.; Joannides, T.; Keily, J. S.;
Laborde, E.; Schroeder, M. C.; Seanie, J. A.; Shapiro, M.
A.; Suto, M. J.; Vanderroest, S. J. Med. Chem. 1993, 36,
871.
8. Suto, M. J.; Domagala, J. M.; Roland, G. E.; Mailloux,
G. B.; Cohen, M. A. J. Med. Chem. 1992, 35, 4745.
9. Jones, M. E.; Critchley, I. A.; Karlowsky, J. A.; Blosser-
Middleton, R. S.; Schmitz, F.-J.; Thornsberry, C.; Sahm,
D. F. Antimicrob. Agents Chemother. 2002, 46, 1651.
10. Yoshiyasu, T.; Hiromi, K.; Nobuyuki, I.; Kazuo, A.
Chem. Pharm. Bull. 1985, 33, 2762.
Substitution on the nitrogen of the side chain also brought
about improvements in the clonogenicity. N-Methylation
showed a dramatic increase in clonogenicity IC50 from
24 lg/mL up to 139 lg/ml in 20 vs 22, respectively,
although the antibacterial activity did show a modest de-
crease. In the azetidine series, the N-substitution of the
ethylcyano group showed a pronounced increase in clo-
nogenicity IC50 (e.g., 34 vs 35 and 36 vs 37), although little
or no effect was seen by the same substitution in the pyr-
rolidine series (e.g., 23 vs 24 and 28 vs 29).
Selected compounds which met the criteria for antibac-
terial activity, dofetilide, and/or clonogenicity were test-
ed in vivo in rats to assess their pharmacokinetic
performance, and the results are shown in Table 2.
The nitrile-containing fluoroquinolones showed good
overall pharmacokinetics. In the C-substituted series,
the pyrrolidine analog 22 (designated PF-00951966)
showed both good bioavailability and AUC. The azeti-
dine analog 37 (designated PF-02298732) showed excel-
lent bioavailability and a good AUC. On the other
hand, the R6 des-fluoro analog (27) showed a sub-opti-
mal volume of distribution and low bioavailability. The
non-nitrile-containing fluoroquinolone, 34, shown for
comparison, showed higher clearance and lower bio-
availability compared to the other compounds tested.
11. S. pneumo.R MICs for stereoisomers of 22: diastereomer A
(faster eluting) = 8 lg/mL, diastereomer B (slower elut-
ing) = 1 lg/mL, and enantiomer B1 (faster eluting) = 8 lg/
mL, enantiomer B2 (slower eluting, desired) = 0.5 lg/mL.
Diastereomers were separated on silica gel eluted with
hexanes/ethyl acetate. Enantiomers were separated by
chiral HPLC on a ChiralPak AD column eluted with
methanol/ethanol.
12. Baldwin, J. E.; Adlington, R. M.; Gollins, D. W.
Tetrahedron 1995, 51, 5169.
13. Frigola, J.; Pares, J.; Corbera, J.; Vano, D.; Merce, R.;
Torrens, A.; Mas, J.; Valenti, E. J. Med. Chem. 1993, 36,
801.
14. Denmark, S. E.; Forbes, D. C.; Hays, D. S.; DePue, J. S.;
Wilde, R. G. J. Org. Chem. 1995, 60, 1391.
15. Unfortunately, these compounds are not very active
against the Enterobacteriaceae. Other than the fastidious
Gram-negatives (H. influenzae and M. catarrhalis) these
compounds have no real efficacy against Gram-negative
In summary, we have synthesized a set of nitrile-con-
taining fluoroquinolones with activity against resistant
strains of S. aureus and S. pneuomoniae and low dis-
placement of dofetilide in the dofetilide binding assay.

S. T. Murphy et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2150–2155
2155
organisms. For example, the MIC50 for compound 22
against Escherichia coli strains and ESBL producing
strains of Klebsiella pneumoniae is 32 lg/ml.
19. Fraley, M. E.; Arrington, K. L.; Buser, C. A.; Ciecko, P.
A.; Coll, K. E.; Fernandes, C.; Hartman, G. D.; Hoffman,
W. F.; Lynch, J. L.; McFall, R. C.; Rickert, K.; Singh, R.;
Smith, S.; Thomas, K. A.; Wong, B. K. Bioorg. Med.
Chem. Lett. 2004, 14, 351.
20. The pKa was measured of the conjugate acid of the most
basic group by capillary electrophoresis: Jia, Z.; Ramstad,
T.; Zhong, M. Electrophoresis 2001, 22, 1112.
21. Gracheck, S. J.; Mychalonka, M.; Gambino, L.; Cohen,
M.; Roland, G.; Ciaravino, V.; Worth, D.; Theiss, J. C.;
Heifetz, C. 31st Interscience Conference on Antimicrobial
Agents and Chemotherapy, Chicago, IL, 1991; Abstract
No. 1488.
16. Compounds in the azetidine series with R = n-propyl,
isopropyl, and cyclopropropyl were also active, but did
not offer any advantages over R = ethyl (37).
17. A filter binding assay with [3H]-dofetilide was used as
described in: Greengrass, P. M.; Stewart, M.; Wood, C.
M. World Patent WO2003021271, 2003.
18. Dofetilide was found to be a useful indicator of hERG
activity for our series. Internally, we have demonstrated a
correlation between the dofetilide displacement assay and
the hERG patch clamp assay for these compounds.