Identification and suppression of a dimer impurity in the development of delafloxacin

Article abstract of DOI:10.1021/op800238q

Delafloxacin is a 6-fluoroquinolone antibiotic which is under development at Rib-X Pharmaceuticals. During initial scale-up runs to prepare delafloxacin, up to 0.43% of a new impurity arose in the penultimate chlorination step. This was identified as a di

Full text of DOI:10.1021/op800238q

Organic Process Research & Development 2009, 13, 54–59
Identification and Suppression of a Dimer Impurity in the Development of
Delafloxacin
Roger Hanselmann,*,† Graham Johnson,† Maxwell M. Reeve,† and Shu-Ting Huang‡
Rib-X Pharmaceuticals, Inc., 300 George Street, Suite 301, New HaVen, Connecticut, 06511, U.S.A., and ScinoPharm Taiwan,
Ltd., No.1, Nan-Ke 8th Road, Tainan Science-Based Industrial Park, Shan-Hua, Tainan County, 74144 Taiwan, R.O.C.
Abstract:
The synthesis of delafloxacin initially underwent develop-
ment by Abbott Laboratories,6 and a key step in this is a
selective chlorination on the 8-position of the functionalized
quinolone 1 (Scheme 1). In this process a solution of 1 in a
mixture of methyl acetate and ethyl acetate is chlorinated using
NCS in the presence of 3.5 mol % of H2SO4, affording 2. This
is followed by solvent exchange and saponification with KOH
to yield 3. Delafloxacin is obtained after salt formation with
N-methyl-D-glucamine.
Despite initial success in implementing this process, we
encountered difficulties upon scale-up of this step in that up to
0.43 area % of a new impurity was detected by HPLC at relative
retention time (RRT) 1.60 after isolation of 3. Additionally, this
new impurity turned out to be difficult to purge during the final
salt formation. Thus, we decided to initiate a study to identify
this impurity, understand how it was being formed, and suppress
its generation.
Delafloxacin is a 6-fluoroquinolone antibiotic which is under
development at Rib-X Pharmaceuticals. During initial scale-up
runs to prepare delafloxacin, up to 0.43% of a new impurity arose
in the penultimate chlorination step. This was identified as a
dimeric adduct of delafloxacin. Subsequent application of design
of experiments (DoE) led to the identification of the factors
responsible for this impurity. Implementation of the knowledge
gained from the DoE reproducibly enabled the suppression of this
impurity to acceptable levels.
Introduction
Antimicrobial resistance in the community and hospital
settings has been a growing public health concern due to the
continuing emergence of multidrug resistant bacterial strains.1
Methicillin-resistant Staphylococcus aureus (MRSA) ranks as
the most frequently isolated pathogen in hospital intensive care
units in the United States where the proportion of MRSA
occurrence in S. aureus isolates increased from 35.9% in 1992
to 64.4% in 2003.2
Since the introduction of nalidixic acid nearly 40 years ago,3
quinolone antibiotics have occupied a prominent place in the
array of antibiotics. 6-Fluoroquinolones such as ciprofloxacin
have especially gained an expanding role in the treatment of
infections, due to their broad spectrum of application.4
Delafloxacin is a 6-fluoroquinolone antibiotic5 with excellent
antibacterial activity against Gram-positive organisms, including
both methicillin-susceptible S. aureus and MRSA. It is currently
undergoing phase II clinical trials.
Results and Discussion
Despite numerous attempts to isolate this new impurity by
preparative HPLC, we were unable to do so, and only the
molecular weight was established by HPLC-MS as 880 Da.
The measured molecular weight of this impurity is exactly
double that of the acid 3, which suggested a dimeric derivative
of this compound. Close examination of the purity profile of
the substrate of the chlorination reaction did not lead to the
detection of any impurity that could similarly be assigned a
dimeric structure, and its formation was therefore attributed to
the chlorination-hydrolysis sequence. A number of potential
dimeric adducts that could arise in this step were considered,
including 4, which would result from cleavage of the azetidine
moiety in one unit of 3 and reaction with the hydroxyl group
of a second. In order to investigate this possibility further, we
embarked on the synthesis of 4.
* To whom correspondence should be addressed. E-mail:rhanselmann@
Rib-x.com.
Retrosynthetically (Scheme 2), molecule 4 is readily discon-
nected into amino alcohol 5 and quinolone 6; the latter is a
known compound.7 A suitably protected amino alcohol 10 was
synthesized from commercially available azetidin-3-ol hydro-
† Rib-X Pharmaceuticals, Inc.
‡ ScinoPharm Taiwan, Ltd.
(1) (a) Cosgrove, S. E.; Carmeli, Y. Clin. Infect. Dis. 2003, 36, 1433. (b)
Seybold, U.; Kourbatova, E. V.; Johnson, J. G.; Halvosa, S. J.; Wang,
Y. F.; King, M. D.; Ray, S. M.; Blumberg, H. M. Clin. Infect. Dis.
2006, 42, 647. (c) Tenover, F. C.; McDougal, L. K.; Goering, R. V.;
Killgore, G.; Projan, S. J.; Patel, J. B.; Dunman, P. M. J. Clin.
Microbiol. 2006, 44, 108.
(6) (a) Haight, A. R.; Ariman, S. Z.; Barnes, D. M.; Benz, N. J.; Gueffier,
F. X.; Henry, R. F.; Hsu, M. C.; Lee, E. C.; Morin, L.; Pearl, K. B.;
Peterson, M. J.; Plata, D. J.; Willcox, D. R. Org. Process Res. DeV.
2006, 4, 751. (b) Barnes, D. M.; Christesen, A. C.; Engstrom, K. M.;
Haight, A. R.; Hsu, M. C.; Lee, E. C.; Peterson, M. J.; Plata, D. J.;
Raje, P. S.; Stoner, E. J.; Tedrow, J. S.; Wagaw, S. Org. Process Res.
DeV. 2006, 4, 803.
(2) Klevens, R. M.; Edwards, J. R.; Tenover, F. C.; McDonald, L. C.;
Horan, T.; Gaynes, R. Clin. Infect. Dis. 2006, 42, 389.
(3) Bush, K. Clin. Microbiol. Infect. 2004, 10 (Suppl. 4), 10.
(4) Emmerson, A. M.; Jones, A. M. J. Antimicrob. Chemother. 2003, 51
(Suppl. S1), 13.
(7) (a) Yazaki, A.; Niino, Y.; Ohshita, Y.; Hirao, Y.; Amano, H.; Hayashi,
N.; Kuramoto, Y. PCT Int. Appl. WO 9711068, 1997. CAN: 126,
305587. (b) Yazaki, A.; Aoki, S. PCT Int. Appl. WO 2001034595,
2001. CAN: 134, 366811.
(5) Delafloxacin was initially developed by Wakunaga Pharmaceuticals
and Abbott Laboratories and was subsequently licensed by Rib-X
Pharmaceuticals, Inc.
54
•
Vol. 13, No. 1, 2009 / Organic Process Research & Development
10.1021/op800238q CCC: $40.75
2009 American Chemical Society
Published on Web 11/26/2008

Scheme 1. Synthesis of delafloxacin
Scheme 2. Retrosynthesis of proposed impurity 4
chloride 7 (Scheme 3). Thus, amino alcohol 10 was condensed
with quinolone 6 to form 11, which after deprotection and a
second condensation with 6 resulted in formation of the dimeric
compound 13. After saponification the proposed impurity 4 was
obtained in 45% overall yield.
With synthetic 4 at hand, the unknown impurity in a
contaminated batch of delafloxacin was compared with syn-
thesized 4 Via spiking experiments and comparison by HPLC-
MS and HPLC-UV. To our delight, synthetic 4 matched
unambiguously with the unknown impurity seen in previously
manufactured batches of delafloxacin.
In order to understand the dynamics of the formation of
impurity 4, we decided to investigate the chlorination reaction
further in a design of experiments (DoE) study. The following
factors were chosen to be investigated in a DoE study of
resolution IV, over ranges specified as follows:8 temperature
(15-25 °C), amount of NCS (1.05-1.2 equiv), amount of
H2SO4 (2-5 mol %), water content in solvent (0-0.5%),
solvent volume (2-3 vol), solvent (methyl acetate/ethyl acetate),
and NCS addition rate (0.05-0.3 vol/min). A total of 19
chlorination reactions were performed in a MultiMax reactor.9
In each case samples from the reactions were quenched after
5 h and saponified with KOH, and then the crude reaction
mixtures were analyzed by HPLC.10 The area % value for
impurity 4 that resulted in each case was processed and analyzed
using the DoE software.11
An excellent correlation of R2 of 0.997 was obtained
following processing of the data. Of the main effects, higher
amounts of NCS, lower temperature, and faster addition of the
NCS solution as well as the use of dry solvents12 had the most
beneficial impact on suppressing the amount of impurity 4
(Figure 1).13 Additionally, a strong interaction was observed
between the amount of NCS and solvent, in that methyl acetate
is preferred when only a slight excess of NCS is used. In order
to suppress any overchlorination of 2, 1.05 equiv of NCS is
preferred, and hence, methyl acetate was chosen as the solvent
of choice for this step.
From a mechanistic point of view, we postulate that impurity
4 could arise from an initial acid-catalyzed activation of the
azetidine ring which triggers an isobutyric ester/chloride-induced
ring opening sequence to 16. During the subsequent saponifica-
tion 16 reacts with the hydrolysis intermediate 17 or 3 to 4
(10) In order to determine the amount of 4, the chlorinated samples 2 were
saponified to 3.
(11) The experimental design and analysis were conducted with JMP,
Design of Experiments, Version 7, SAS Institute Inc., Cary, NC,
U.S.A., 1989-2007, using a stepwise fit followed by a standard least
squares method.
(8) The experimental table and results can be found in the Supporting
(12) Methyl acetate containing less than 500 ppm water was used, prior to
adjustment as required by the appropriate experiment in the DoE.
(13) A more detailed analysis can be found in the Supporting Inform-
ation.
Information.
(9) Available from Mettler-Toledo, Inc., 1900 Polaris Parkway, Columbus,
OH 43240, U.S.A.
Vol. 13, No. 1, 2009 / Organic Process Research & Development
•
55

Figure 1. Prediction profiler for ethyl acetate as the solvent.
Scheme 3. Synthesis of impurity 4
14
(Scheme 4).
The validity of this sequence was further
Thus, an acceptable turnaround time for the in-process control
can be achieved when a level of 1% of H2SO4 is employed as
catalyst.
strengthened by a subsequent HPLC-MS analysis of a crude
chlorination reaction before saponification. In this, an impurity
with a molecular weight of 574 Da, which matches 16, was
detected in approximate equal amounts compared to 4 after
saponification.
After having established a good understanding of the critical
parameters with respect to formation of this impurity, a second
DoE study was initiated to test the robustness of the reaction
in the anticipated process operating range. In this, a DoE study
of resolution IV was designed with the following factors
undergoing variation: temperature (13-21 °C), amount of NCS
(1.04-1.07 equiv), NCS addition rate (30-75 min), and H2SO4
(0.8-1.2 mol %). A total of 10 chlorination reactions were
performed in a MultiMax reactor. In each case samples were
quenched and saponified after passing the in-process control.
The resulting area % of 4 was processed and analyzed using
the DoE software. As anticipated, temperature, amount of NCS,
and H2SO4 had a statistically significant effect on the amount
of impurity 4 in the studied parameter range. However,
assuming the worst case scenario in the prediction profiler,
impurity 4 has a value of 0.11 area % ( 0.01%, which is well
within the acceptable limit that has been established from a
toxicological batch of delafloxacin (Figure 3).
Based on this hypothetical mechanism, a time dependency
for the formation of 4 during the chlorination process cannot
be excluded, and since the reaction time was kept constant in
the DoE study, it was decided to evaluate this parameter
independently. A chlorination reaction was performed using
3.5% of H2SO4 and methyl acetate as solvent at 15 °C, and a
sample was quenched after the reaction was deemed to be
complete. Additional samples were quenched after 2 h and 6 h,
saponified, and analyzed by HPLC. Not surprisingly, a steady
increase of the impurity 4 over time was seen. This result has
an impact on controlling the chlorination process, in that an
adequate turnaround time for the HPLC monitoring of this
reaction would be necessary in order to minimize the formation
of 4. However, subsequent experiments showed that decreasing
the amount of H2SO4 to 1% diminished the amount of impurity
4 produced over time without having a significant impact on
the chlorination reaction time or the quality of 3 (Figure 2).
Two subsequent kilo laboratory runs of this reaction
confirmed the effectiveness of the parameter changes, and
material of high quality with impurity 4 levels of 0.07% were
obtained after saponification.
(14) Saponification and subsequent epoxide formation of 16 prior to
condensation with 3 or 17 cannot be ruled out.
56
•
Vol. 13, No. 1, 2009 / Organic Process Research & Development

Scheme 4. Proposed mechanism of impurity 4
and the solution was seeded with 3 (27 g, 61 mmol). The
suspension was stirred for 1 h at 40 °C, and then aqueous acetic
acid (13%, 11.7 kg) was slowly added. After stirring an
additional hour at 40 °C, the suspension was cooled to room
temperature, filtered, washed with water (41 kg), and dried at
60 °C/vacuum to yield 3 as yellow crystals (2.5 kg, 91%).
Isolated 3 had the same spectroscopic properties as reported.6
3-Hydroxyazetidine-1-carboxylic Acid Benzyl Ester, 8.
To a solution of azetidin-3-ol hydrochloride 7 (25 g, 0.23 mol)
in water (150 mL) and THF (300 mL) was added K2CO3 (63.1
g, 0.46 mol). The mixture was stirred for 30 min. at 20-25
°C. Then benzyl chloroformate (40.9 g, 0.24 mol) was added
within 30 min. at 0-5 °C followed by stirring the mixture
overnight at 20-25 °C. THF was removed on a rotavap at 30
°C/vacuum and the mixture was extracted with ethyl acetate
(2 × 150 mL). The combined organic layer was washed with
water (1 × 50 mL), dried over Na2SO4 and concentrated. The
residue was purified by flash column chromatography on silica
gel, eluting with ethyl acetate-heptanes 1:1 and 4:1 to yield 8
as a clear oil (47.3 g, 100%). 1H NMR (300 MHz, CDCl3): δ
3.72 (1H, d, J ) 6.2 Hz), 3.85 (2H, dd, J ) 9.5, 4.4 Hz), 4.17
(2H, dd, J ) 9.5, 6.7 Hz), 4.49-4.57 (1H, m), 5.06 (2H, s),
7.31-7.38 (5H, m); 13C NMR (75 MHz, CDCl3): δ 59.2, 61.6,
66.9, 127.9, 128.1, 128.5, 136.5, 156.6; IR: (film) 3406, 1686,
1438 cm-1; ES-HRMS m/z: (M+ + 1H) calcd for C11H14NO3
208.0968, found 208.0967.
Figure 2. Impact of H2SO4 on impurity 4 levels.
In conclusion, we have successfully identified a dimer
impurity which was detected during scale-up of delafloxacin.
Subsequent DoE experiments enabled us to identify the means
to control this impurity to acceptable levels in small scale as
well as in kilo laboratory runs.
Experimental Section
1-(6-Amino-3,5-difluoropyridin-2-yl)-8-chloro-6-fluoro-7-
(3-hydroxyazetidin-1-yl)-4-oxo-1,4-dihydroquinoline-3-car-
boxylic Acid, 3, Improved Procedure. To a suspension of 1
(3.1 kg, 6.15 mol) in methyl acetate (8.6 kg) was added a
solution of H2SO4 (5.9 g, 62 mmol) and NCS (0.88 kg, 6.46
mol) in methyl acetate (14.4 kg) at 10-17 °C within 45 min.
The solution was stirred at 13-19 °C for 2 h and quenched
with 1.6% aqueous NaHCO3 (12.6 kg), and the organic layer
was washed with 11% aqueous Na2SO3 (7 kg). The methyl
acetate solution was solvent exchanged to 2-propanol at 50 °C/
vacuum; then a solution of KOH (1.1 kg, 19.7 mol) in water
(24.8 kg) was added, and the mixture was stirred at 55 °C for
3 h. Aqueous acetic acid (13%, 2.6 kg) was added at 40 °C,
3-Oxiranylmethoxyazetidine-1-carboxylic Acid Benzyl
Ester, 9. To a solution of 8 (30 g, 0.15 mol) in DMSO (250
mL) was slowly added a solution of NaOH (9.9 g, 0.25 mol)
in water (195 mL) at 15-25 °C. Epichlorohydrin (93.8 g, 1.01
mol) was added, and the mixture was stirred at 20-25 °C for
Vol. 13, No. 1, 2009 / Organic Process Research & Development
•
57

Figure 3. Worst case scenario in prediction profiler for robustness DoE.
24 h. The mixture was diluted with water (300 mL) and
extracted with ethyl acetate (2 × 150 mL). The combined
organic layer was washed with water (2 × 50 mL), dried over
Na2SO4, and concentrated. The residue was purified by flash
column chromatography on silica gel, eluting with ethyl acetate/
heptanes, 3:2, yielding 9 as a clear oil (32.1 g, 84%). 1H NMR
(300 MHz, CDCl3): δ 2.60 (1H, dd, J ) 4.8, 2.6 Hz), 2.81
(1H, dd, J ) 4.9, 4.2 Hz), 3.09-3.16 (1H, m), 3.25 (1H, dd, J
) 11.4, 6.2 Hz), 3.68 (1H, dd, J ) 11.5, 2.5 Hz), 3.89-3.97
(2H, m), 4.15-4.24 (2H, m), 4.29-4.37 (1H, m), 5.09 (2H, s),
7.28-7.36 (5H, m); 13C NMR (75 MHz, CDCl3): δ 44.2, 50.4,
56.7, 56.9, 66.7, 68.6, 70.0, 128.0, 128.1, 128.5, 136.6, 156.5;
IR: (film) 2951, 1709, 1420 cm-1; ES-HRMS m/z: (M+ + 1H)
calcd for C14H18NO4 264.1230, found 264.1230.
roquinoline-3-carboxylic acid ethyl ester, 12. To a slurry of
10% Pd on carbon (2.1 g) in MeOH (20 mL) was added a
solution of 11 (13.7 g, 20.3 mmol) in MeOH (230 mL). The
mixture was hydrogenated at 1 atm for 1 h, filtered over Hyflo,
and evaporated yielding 12 as beige crystals (10.3 g, 93%). Mp
148-152 °C; 1H NMR (300 MHz, DMSO-d6): δ 1.27 (3H, t,
J ) 7.1 Hz), 3.27 (1H, d, J ) 5.0 Hz), 3.28-3.80 (10H, m),
4.19 (1H, br s), 4.21 (2H, q, J ) 7.1 Hz), 5.86 (1H, s), 6.74
(2H, s), 7.84 (1H, d, J ) 13.8 Hz), 7.94 (1H, dd, J ) 9.7, 9.0
Hz), 8.43 (1H, s); 13C NMR (75 MHz, CDCl3): δ 14.1, 48.4
(d, JF ) 10 Hz), 53.6, 60.2, 68.4, 70.4 (d, JF ) 4 Hz), 72.1,
106.4 (d, JF ) 6 Hz), 111.0, 111.3 (d, JF ) 23 Hz), 113.6 (dd,
JF ) 23, 21 Hz), 118.9 (d, JF ) 6 Hz), 133.8 (d, JF ) 13 Hz),
134.2, 139.5 (d, JF ) 12 Hz), 143.3 (dd, JF ) 248, 4 Hz), 145.0
(dd, JF ) 259, 5 Hz), 145.6 (d, JF ) 14 Hz), 149.3 (d, JF )
245 Hz), 149.5, 163.5, 171.0; IR: (KBr) 1697, 1614, 1496, 1457
cm-1; ES-HRMS m/z: (M+ + 1H) calcd for C23H24ClF3N5O5
542.1413, found 542.1391.
1-Amino-3-(azetidin-3-yloxy)-propan-2-ol-bis(N,N′-qui-
nolone Diester), 13. A solution of 12 (9.6 g, 17.7 mmol), 6
(7.8 g, 18.6 mmol), and N,N-diisopropylethylamine (4.6 g, 35.4
mmol) in NMP (150 mL) was stirred at 55 °C for 3 h. The
solution was poured into 1 N citric acid/ice (300 mL) and
extracted with ethyl acetate (3 × 100 mL). The combined
organic layers were washed with water (2 × 100 mL), dried
over Na2SO4, and concentrated. The residue was purified by
flash column chromatography on silica gel, eluting with ethyl
acetate/MeOH, 95:5. The obtained yellow foam was crystallized
with CH2Cl2/MeOH, 9:1 (160 mL), yielding 13 as beige crystals
(11.8 g, 71%). Mp 184-187 °C; 1H NMR (300 MHz, DMSO-
d6): δ 1.26 (6H, t, J ) 7.1 Hz), 3.29-3.48 (3H, m), 3.49-3.62
(1H, m), 3.73-3.82 (1H, m), 4.12-4.30 (3H, m), 4.21 (4H, q,
J ) 7.1 Hz), 4.52-4.65 (2H, m), 5.13-5.22 (1H, m),
5.83-5.92 (1H, m), 6.72 (4H, s), 7.73 (1H, d, J ) 13.9 Hz),
7.82 (1H, d, J ) 13.9 Hz), 7.92 (1H, t, J ) 9.6 Hz), 7.93 (1H,
t, J ) 8.7 Hz), 8.41 (2H, s); 13C NMR (75 MHz, CDCl3): δ
12.3 (2×), 46.4 (d, JF ) 11 Hz), 58.4 (2×), 61.9 (2×), 66.7,
67.3 (d, JF ) 4 Hz), 69.1, 103.4 (d, JF ) 6 Hz), 104.6 (d, JF )
6 Hz), 108.7 (d, JF ) 23 Hz), 109.2, 109.4 (d, JF ) 23 Hz),
109.5, 111.7 (dd, JF ) 25, 24 Hz), 111.8 (dd, JF ) 25, 24 Hz),
117.1 (d, JF ) 7 Hz), 117.8 (d, JF ) 6 Hz), 132.1 (dd, JF ) 17,
4 Hz), 132.2, 132.5, 133.5, 137.7 (d, JF ) 12 Hz), 139.4 (d, JF
) 12 Hz), 141.0 (dd, JF ) 247, 5 Hz), 141.5 (dd, JF ) 248, 5
Hz), 143.0 (dd, JF ) 259, 5 Hz), 143.3 (dd, JF ) 259, 5 Hz),
143.8 (2×), d, JF ) 15 Hz), 147.5 (d, JF ) 245 Hz), 147.7,
147.8, 148.1 (d, JF ) 247 Hz), 161.7 (2×), 169.1, 169.2; IR:
1-(6-Amino-3,5-difluoropyridin-2-yl)-7-[3-(1-benzyloxy-
carbonylazetidin-3-yloxy)-2-hydroxypropylamino]-8-chloro-
6-fluoro-4-oxo-1,4-dihydroquinoline-3-carboxylic Acid Ethyl
Ester, 11. A mixture of 9 (19 g, 72.2 mmol) in conc NH4OH
(380 mL) and 7 M NH3 in MeOH (86 mL) was stirred for 5 h
at room temperature. The clear solution was concentrated and
azeotropically dried with toluene. The residual clear oil and 6
(20 g, 48.1 mmol) were dissolved in NMP (150 mL). N,N-
Diisopropylethylamine (12.4 g, 96.2 mmol) was added, and the
solution was stirred at 70 °C for 3 h. The solution was poured
into 1 N citric acid/ice (300 mL) and extracted with ethyl acetate
(2 × 150 mL). The combined organic layers were washed with
water (2 × 100 mL), dried over Na2SO4, and concentrated. The
residue was purified by flash column chromatography on silica
gel, eluting with ethyl acetate/heptanes, 1:1 followed by ethyl
acetate/MeOH, 95:5, yielding 11 as a yellow foam (27.1 g,
1
83%). H NMR (300 MHz, CDCl3): δ 1.35 (3H, t, J ) 7.1
Hz), 3.35-3.52 (4H, m), 3.62-3.77 (1H, m), 3.84-3.91 (2H,
m), 3.95-4.08 (1H, m), 4.15 (2H, dd, J ) 9.3, 6.5 Hz),
4.23-4.30 (1H, m), 4.35 (2H, q, J ) 7.1 Hz), 4.85-5.13 (3H,
br s), 5.08 (2H, s), 7.18-7.25 (1H, m), 7.31-7.35 (5H, m),
7.99 (1H, dd, J ) 13.7, 3.1 Hz), 8.31 (1H, s); 13C NMR (75
MHz, CDCl3): δ 14.4, 48.5 (d, JF ) 10 Hz), 56.6, 61.1, 66.9,
68.6, 69.3, 70.8, 107.2, 111.5, 112.6 (d, JF ) 24 Hz), 113.2
(m), 120.6, 128.0, 128.1, 128.5, 134.1 (d, JF ) 5 Hz), 134.7
(m), 136.5, 139.2 (d, JF ) 13 Hz), 144.9 (d, JF ) 253 Hz),
144.4 (d, JF ) 13 Hz), 145.6 (dd, JF ) 262, 4 Hz), 149.9 (d, JF
) 246 Hz), 150.0, 156.5, 164.7, 172.9; IR: (KBr) 2949, 1700,
1615 cm-1; ES-HRMS m/z: (M+ + 1H) calcd for
C31H30ClF3N5O7 676.1780, found 676.1762.
1-(6-Amino-3,5-difluoropyridin-2-yl)-7-[3-(azetidin-3-yloxy)-
2-hydroxypropylamino]-8-chloro-6-fluoro-4-oxo-1,4-dihyd-
58
•
Vol. 13, No. 1, 2009 / Organic Process Research & Development

(KBr) 1728, 1615, 1491, 1448 cm-1; ES-HRMS m/z: (M+ +
1H) calcd for C40H33Cl2F6N8O8 937.1697, found 937.1696.
1-Amino-3-(azetidin-3-yloxy)-propan-2-ol-bis(N,N′-qui-
nolone Carboxylic Acid), 4. To a suspension of 13 (17.0 g,
18.1 mmol) in 2-propanol (75 mL) was added a 1 N KOH
solution (127 mL, 126.7 mmol). After stirring the mixture at
55 °C for 3.5 h, the solution was cooled to 30 °C, and a solution
of AcOH (12.4 g, 206.5 mmol) dissolved in water (94 mL)
was added within 1 h. The suspension was stirred at room
temperature for 2 h, filtered, washed with water (3 × 40 mL),
and dried at 50 °C/vacuum, yielding 4 as yellow crystals (15.7
g, 98%). Mp 198-205 °C dec; 1H NMR (300 MHz, DMSO-
d6): δ 3.28-3.45 (2H, m), 3.45-3.78 (2H, m), 3.79-3.88 (1H,
m), 4.16-4.33 (3H, m), 4.61-4.75 (2H, m), 5.25 (1H, br s),
6.23-6.35 (1H, m), 6.76 (4H, s), 7.79 (1H, d, J ) 13.7 Hz),
7.90 (1H, d, J ) 13.8 Hz), 7.93 (2H, dd, J ) 9.7, 2.4 Hz), 8.70
(1H, s), 8.71 (1H, s), 14.59 (2H, br s); 13C NMR (75 MHz,
CDCl3): δ 48.1 (d, JF ) 11 Hz), 63.8, 68.4, 69.0 (d, JF ) 5
Hz), 70.6 (d, JF ) 6 Hz), 104.5 (d, JF ) 6 Hz), 105.9 (d, JF )
7 Hz), 107.8, 108.2, 109.8 (d, JF ) 23 Hz), 110.8 (d, JF ) 23
Hz), 113.4 (d, JF ) 23 Hz), 113.7 (d, JF ) 23 Hz), 115.8 (d, JF
) 8 Hz), 116.6 (d, JF ) 8 Hz), 133.3 (dd, JF ) 14, 3 Hz),
133.5 (dd, JF ) 14, 4 Hz), 134.8, 135.9, 141.0 (d, JF ) 12 Hz),
142.1 (d, JF ) 12 Hz), 142.8 (dd, JF ) 249, 5 Hz), 143.3 (dd,
JF ) 249, 5 Hz), 145.1 (dd, JF ) 259, 5 Hz), 145.4 (dd, JF )
260, 5 Hz), 145.6 (2×, d, JF ) 15 Hz), 149.5 (d, JF ) 248
Hz), 150.1 (2×), 150.2 (d, JF ) 249 Hz), 164.7, 164.8, 175.8
(d, JF ) 3 Hz), 175.9 (d, JF ) 3 Hz); IR: (KBr) 1727, 1622,
1489, 1439 cm-1; ES-HRMS m/z: (M+ + 1H) calcd for
C36H25Cl2F6N8O8 881.1071, found 881.1090.
Acknowledgment
We thank Derek Robinson for initial guidance on the Design
of Experiments study and Aihua Kao for experimental assistance.
Supporting Information Available
Experimental tables and analyses of the DoE studies. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Received for review September 25, 2008.
OP800238Q
Vol. 13, No. 1, 2009 / Organic Process Research & Development
•
59