Synthesis and antibacterial activity of dihydro-1,2-oxazine and 2-pyrazoline oxazolidinones: Novel analogs of linezolid

Article abstract of DOI:10.1016/j.bmcl.2005.03.099

The synthesis and antibacterial activity of oxazolidinones containing dihydro-1,2-oxazine and 2-pyrazoline ring systems are described. Linezolid analogs utilizing dihydro-1,2-oxazines as morpholine mimics were prepared utilizing a nitrosoamine/diene 4+2 c

Full text of DOI:10.1016/j.bmcl.2005.03.099

Bioorganic & Medicinal Chemistry Letters 15 (2005) 2834–2839
Synthesis and antibacterial activity of dihydro-1,2-oxazine and
2-pyrazoline oxazolidinones: novel analogs of linezolid
Stan DÕAndrea,* Zhizhen Barbara Zheng, Kenneth DenBleyker, Joan C. Fung-Tomc,
Hyekyung Yang, Junius Clark, Dennis Taylor and Joanne Bronson
Bristol-Myers Squibb, Discovery Chemistry, Pharmaceutical Research Institute, 5 Research Parkway, Wallingford, CT 06492, USA
Received 18 June 2004; revised 22 March 2005; accepted 25 March 2005
Available online 28 April 2005
Abstract—The synthesis and antibacterial activity of oxazolidinones containing dihydro-1,2-oxazine and 2-pyrazoline ring systems
are described. Linezolid analogs utilizing dihydro-1,2-oxazines as morpholine mimics were prepared utilizing a nitrosoamine/diene
4+2 cycloaddition strategy. Pyrazolidine, hexahydro-pyridazine, and 2-pyrazoline analogs more closely related to eperezolid were
also prepared. The most active of these new oxazolidinones were the dihydro-1,2-oxazine 6 and the 2-pyrazoline 20 both of which
had potency similar to linezolid against a panel of Gram-positive bacteria.
ꢀ 2005 Elsevier Ltd. All rights reserved.
The oxazolidinone antibacterial linezolid (1) discovered
and developed by Upjohn/Pharmacia (now marketed by
Pfizer as ZyvoxTM) has been described as the first new
class of antibacterial agents to be marketed in the last
30 years. The related eperezolid (2) (Fig. 1) was not ad-
VRSA2 infections. Another distinguishing feature of this
class of antibacterial agents is its synthetic origin.
Efforts to enhance the spectrum and potency of this class
of antibiotics have been ongoing throughout the phar-
maceutical industry.3 It has been argued that the most
urgent improvement needed for the oxazolidinone class
is an enhanced safety profile.4 Hematological toxicity
(possibly due to myelo-suppression) is a side effect espe-
cially during prolonged treatment (>14 days). Based on
our analysis of the limited published oxazolidinone toxi-
city data available at the time, we concluded that the
most prudent approach would be to maintain the amine
functionality at the 4-position of the phenyl ring.
vanced to Phase 3 studies presumably due to its shorter
1
human T1/2
.
The urgency of discovering new classes of antibacterial
agents was highlighted when the first clinical case of
vancomycin-resistant Staphylococcus aureus (VRSA)
infection was reported in 2002.2 Clinically, one of linezo-
lidÕs most valuable characteristics is that it is one of the
few orally active agents available to treat MRSA1 and
Figure 1. Linezolid and eperezolid.
Keywords: Antibacterial; Oxazolidinones; Linezolid.
*
Corresponding author. Tel.: +1 203 677 6696; fax: +1 203 677 7702; e-mail: stanley.dandrea@bms.com
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.03.099

S. DÕAndrea et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2834–2839
2835
Figure 2. Pseudo-isomers of linezolid and eperezolid.
Scheme 1. Synthesis of hydroxylamine and nitroso derivative.
Reagents and conditions: (a) Zn, NH4Cl, 60 ꢁC, 10 min; (b) PCC, rt,
10 min, THF/CH3CN (5:2).
Further, we felt that modulating (either reducing or increas-
ing) the electron-donating ability of the nitrogen in the
4-position of the phenyl ring could have a beneficial
effect on the safety profile.5 It should be emphasized
that these presumed structure/toxicity relationships re-
main unproven to date. The SAR of amines in the 4-po-
sition was studied in detail by Upjohn/Pharmacia, but
4-amino groups imparting useful properties continue to
emerge.6–8 In this paper, we will detail our efforts to
replace the linezolid morpholine by the pseudo-isomeric
dihydro-1,2-oxazine and isoxazolidine ring systems. In
addition, we will describe the replacement of the eper-
ezolid piperazine by pyrazolidine, hexahydropyridazine,
and 2-pyrazoline rings (Fig. 2).
ence of cyclohexadiene provided the dihydro-1,2-ox-
azine
8 (31%) (Scheme 2). Attempted catalytic
hydrogenation of the olefinic double bond in 6 did not
result in the formation of 9, instead, the N–O bond
was severed and the ring-opened product 10 was formed
(Scheme 3). This result was somewhat surprising based
on the successful catalytic hydrogenation of similar
dihydro-1,2-oxazines.13 Attempted diimide reduction
of 8 resulted in no reaction.
The fluoro analog of 5 was sought as a key intermediate
to prepare a fluorinated series of dihydro-1,2-oxazines.
Nitration of 3-fluoro-oxazolidinone 1114 provided a
1:1 ratio of the 6-nitro oxazolidinone 12 and the desired
4-nitro regioisomer 13 (Scheme 4). Separation by pre-
parative HPLC gave a 20% yield of each pure
These morpholine and piperazine replacements were
chosen based on their modest structural resemblance,
their modified electronic properties,9 and the fact that
they remained unexplored by previous researchers. In
general, these cyclic amines containing an N–O or N–N
linkage are less basic and less electron donating than
the corresponding morpholine, piperazine, and pyrroli-
dine ring systems.10 Inspired by a synthetic strategy used
to functionalize the 7-position of quinolone antibacteri-
als,11 the known nitro oxazolidinone 312 was reduced
with Zn/NH4Cl to give the hydroxylamine 4 in 94%
yield. Oxidation of 4 with PCC provided the nitroso
derivative 5 in 56% isolated yield (Scheme 1). The
[4+2] cycloaddition of 5 with butadiene and 2,3-dimeth-
ylbuta-1,3-diene gave rise to dihydro-1,2-oxazines 6 and
7 (89% and 57% yield, respectively) (Scheme 2). In some
cases, it was possible to carry out this two step sequence
in one pot by generating the nitroso derivative 5 in situ
in the presence of the appropriate diene. For instance,
this in situ oxidation procedure carried out in the pres-
Scheme 3. Reduction of dihydro-1,2-oxazine 6 (H2, 5% Pd/C, EtOAc,
Parr shaker, 2 h).
Scheme 2. Nitroso amine/diene 4+2 cycloadditions. Reagents and conditions: (a) butadiene or 2,3-dimethylbuta-1,3-diene, 0 ꢁC to rt, 3 h, CH2Cl2;
(b) 4.3 equiv Bu4NIO4, cyclohexadiene, DMF, rt, 3 h.

2836
S. DÕAndrea et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2834–2839
to selectively reduce the nitro group in 15 failed (Pd/
NH4OCHO, SnCl2, or Na2S2O4).
The 2-pyrazoline 20 was prepared as shown in Scheme 6.
Hydrazine was added to 3,4-difluoronitrobenzene to
give a 98% yield of arylhydrazine 16. Dialkylation of
the hydrazine moiety with 1,3-dibromopropane fol-
lowed by air oxidation led to a 39% yield of 2-pyrazoline
17. Optimization of these reaction conditions was not
undertaken, so it remains uncertain whether other
oxidants would be superior to air for this conversion.
Under the air oxidation conditions employed, no sign
of over oxidation to the pyrazole was observed although
minor impurities were not investigated. Phase transfer
catalytic hydrogenation provided the aniline (84%
yield), which was treated with benzylchloroformate to
give the carbamate 18 (80% yield). Note that these are
the same catalytic phase transfer reduction conditions
which led to over-reduction (N–O bond cleavage) in
the above case (see Scheme 5). At this point, general
methodology documented by Upjohn/Pharmacia1 was
employed. The anion of carbamate 18, formed using
n-BuLi, was quenched by the addition of (R)-glycidyl
butyrate. The hydroxymethyl oxazolidinone 19 is ob-
tained upon workup (75% yield). Mesylate formation
followed by aminolysis and acetylation ultimately gave
the final product 20 (50% yield over three steps).
Scheme 4. Nitration of 11. Reagents and conditions: (a) KNO3,
H2SO4, 0 ꢁC to rt, 4 h.
Scheme 5. Isoxazolidine analog.
regioisomer. Unfortunately, we were unsuccessful in our
attempt to reduce 13 to the hydroxylamine. This pre-
vented our synthesis of the fluorinated dihydro-1,2-
oxazine series. Alternate, more laborious routes to this
series were considered but regrettably never undertaken.
An approach to the five-membered ring analog of 9 is
shown in Scheme 5. Isoxazolidine 14 reacted smoothly
with 4-fluoronitrobenzene to give isoxazolidine 15
(86%). While selective reduction of the nitro group
seemed to be a risky proposition, we gained some confi-
dence after finding precedence for the selective reduction
of a Cbz group,15 a urea,16 and an oxime moiety17 all in
the presence of a N–O bond. Unfortunately, all attempts
Analogs 25 and 26, which are related to eperezolid, were
prepared similarly as shown in Scheme 7. The bis TFA
salts of pyrazolidine and hexahydro-pyridazine18 were
reacted with 3,4-difluoronitrobenzene in the presence
of excess HunigÕs base to give nitro derivatives 21 and
¨
22 (30% and 83% yield, respectively). Acylation pro-
vided N-acetate derivatives 23 and 24, which were
Scheme 6. Synthesis of 2-pyrazoline analog 20. Reagents and conditions: (a) hydrazine hydrate, pyridine, rt, 50 min; (b) NaH, 1,3-dibromopropane,
DMF, 0 ꢁC to rt, 3 h; (c) air; (d) 10% Pd/C, HCO2NH4, THF/MeOH (1:2); (e) CbzCl, NaHCO3, H2O/acetone (1:2), 0 ꢁC to rt, 2 h; (f) n-BuLi, (R)-
glycidyl butyrate, THF, À78 ꢁC to rt, 18 h; (g) MsCl, NEt3, CH2Cl2, rt, 1 h; (h) NH4OH, THF, iPrOH, 100 ꢁC, 5 h, sealed tube; (i) Ac2O,
triethylamine, CH2Cl2, rt, 2 h.

S. DÕAndrea et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2834–2839
2837
converted to oxazolidinones 25 and 26 via standard
methods.1
The minimum inhibitory concentrations19 (MICs) are
shown in Table 1. This panel is composed of some rep-
resentative Gram-positive bacteria, including strepto-
cocci, staphylococci, and enterococci. It should be
noted that the staph aureus strain is methicillin-suscepti-
ble and penicillinase positive, and the enterococci are
not vancomycin-resistant strains. The dihydro-1,2-oxa-
zine series appeared to be sensitive to the overall size
of the oxazine ring. The smallest, least substituted mem-
ber 6 showed the best potency against this panel of
Gram-positive bacteria, while the most bulky dihydro-
1,2-oxazine 8 was the least potent. This is consistent
with established oxazolidinone SAR indicating that
excessive steric bulk adjacent to the aryl ring is detri-
mental to activity.20 Oxazolidinone 6 is roughly equiva-
lent in potency versus linezolid against this panel. It is
Scheme 7. Synthesis of 25 and 26. Reagents and conditions: (a)
iPr2NEt, and pyrazolidine or hexahydro-pyridazine, CH3CN, 90 ꢁC,
10 h; (b) Ac2O, pyridine, DMAP, THF, rt, 1 h.
Table 1. Minimum inhibitory concentrations (lg/mL)
Compd
X
Y
Minimum inhibitory concentration (lg/mL)a
S. aureusb
S. pneumoc
MRSAd
MRSAe
E. faecalisf
E. faeciumg
1 Linezolid
2 Eperezolid
F
Morpholine
HOCH2C(O)piperazine
1
1
0.5
16
2
2
2
4
2
F
1
1
1
2
128
4
2
128
4
Ref.
3
H
H
F
NH2
NO2
NO2
NHOH
NO
16
0.5
>128
0.5
2
128
2
64
1
13
4
64
4
128
4
128
2
>128
8
>128
4
H
H
5
1
2
1
4
2
6
7
H
H
1(1)
4
2
8
2(2)
8
1(1)
4
4
2
8
16
8
H
F
16
32
16
8
4
32
32
16
32
25
0.5
32
16
26
20
F
F
32
4
1
32
16
32
2
32
4
1(1)
1(1)
1(1)
a Values in parentheses are MICÕs in the presence of 50% calf serum.
b Staphylococcus aureus/Pen+ A15090.
c Streptococcus pneumoniae A9585.
d Staphylococcus aureus/heterogeneously methicillin-resistant A27218.
e Staphylococcus aureus/homogeneously methicillin-resistant A27223.
f Enterococcus faecalis A20688.
g Enterococcus faecium A24885.

2838
S. DÕAndrea et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2834–2839
2. Charles Stratton, MD. VRSA: The worst has finally
happened. 42nd Intersci. Conf. Antimicrob Agents Che-
mother (ICAAC), San Diego, CA, September 27, 2002.
3. AZD2563 is an oxazolidinone in early development at
AstraZeneca. RBX-7644 (ranbezolid) is an oxazolidinone
in early development at Ranbaxy. See: Ednie, L. M.;
Rattan, A.; Jacobs, M. R.; Appelbaum, P. C. Antimicrob.
Agents Chemother. 2003, 47, 1143–1147.
quite possible that the fluoro analog of 6 would
have been more active than linezolid against this panel
of Gram-positive bacteria. This ÔfluoroÕ effect is
evident in most of the oxazolidinone series (e.g., piper-
azines) in which sufficient MIC data have been
published.1
Turning to the eperezolid analogs, the pyrazolidine
derivative 25 and the hexahydro-pyridazine 26 had mod-
erate to poor antibacterial activity.21 It is possible that
the potency of these derivatives could be enhanced by
more closely mimicking the eperezolid SAR and deriva-
tizing the pyrazolidine and hexahydropyridazine rings
with the 2-hydroxyacetyl group in place of the acetate.1
One can speculate further on why compounds 25 and 26
did not maintain antibacterial potency on par with eper-
ezolid. Perhaps acylation of the pyrazolidine and hexa-
hydropyridazine rings diminishes the electron-donating
ability of these heteroatom ring systems relative to an
acylated piperazine. Since the chemical space occupied
by the acetyl group in 25 and 26 is quite distinct from
eperezolid, itÕs possible that steric interaction of the ace-
tate forces the aryl-nitrogen bond to rotate out of the
plane. The 2-pyrazoline analog 20 had potency similar
to linezolid and eperezolid. In fact, 20 appears to have
slightly lower MICs than linezolid against MRSA and
Enterococcus faecalis. Some of the oxazolidinone inter-
mediates were also assayed for in vitro antibacterial
activity. Interestingly, hydroxylamine 4 was significantly
more potent than the corresponding aniline.22 The two
nitro derivatives 3 and 13 represent a notable exception
to the Ôfluoro effectÕ mentioned above. An alternate
explanation is that the highly electrophilic ortho-fluo-
ronitrophenyl moiety is insoluble, or unstable under
the assay conditions.
4. Ethan Rubinstein, MD. 43rd Intersci. Conf. Antimicrob
Agents Chemother (ICAAC), September 14–17, 2003,
Session 176(F,L) abstract#1782. (The need for an
improved hematological profile is less clear cut for routine
therapy: analysis of safety assessments on 2046 Zyvox
patients and 2001 comparator drug-treated patients from
seven Phase III clinical trials indicated virtually no
differences in effects on hemoglobin, neutrophil, or platelet
counts over the first 14 days of therapy. See: Rubinstein,
E.; Isturiz, R.; Standiford, H. C.; Smith, L. G.; Oliphant,
T. H.; Cammarata, S.; Hafkin, B.; Le, V.; Remington, J.
Antimicrob. Agents Chemother. 2003, 47, 1824–1831).
5. Most of the oxazolidinones reported by Upjohn/Pharma-
cia to have superior safety profiles in a 30-day rat toxicity
model have donating nitrogens in the para-position of the
phenyl group. These agents include linezolid, eperezolid,
U-100480 (thiomorpholine), and U-97456 (indoline). See
Ref. 1. DuPont agents DuP-721 and DuP-105 lacking a
para-amine functionality had more severe toxicity in a 30-
day rat model (in house data).
6. Weidner-Wells, M. A.; Boggs, C. M.; Foleno, B. D.;
Melton, J.; Bush, K.; Goldschmidt, R. M.; Hlasta, D.
Bioorg. Med. Chem. Lett. 2003, 13, 4173–4177.
7. Kim, H. Y.; Lee, J. S.; Cha, J. H.; Pae, A. N.; Cho, Y. S.;
Chang, M. H.; Koh, H. Y. Bioorg. Med. Chem. Lett. 2003,
13, 2227–2230.
8. Fukuda, Y.; Hammond, M. L. PCT. WO2003027083-A1,
April 03, 2003.
9. For a discussion of 3D-QSAR of 3-aryloxazolidinone
antibacterials see: Karki, R. G.; Kulkarni, V. M. Bioorg.
Med. Chem. 2001, 9, 3153–3160.
10. Coutts, R. T.; El-Hawari, A. Heterocycles 1974, 2, 669–
743.
11. Ziegler, C. B., Jr.; Bitha, P.; Lin, Y. J. Heterocycl. Chem.
1988, 25, 719.
12. Park, C. H.; Britelli, D. R.; Wang, C. L. J.; Marsh, F. D.;
Gregory, W. A.; Wuonola, M. A.; McRipley, R. J.;
Eberly, V. S.; Slee, A. M.; Forbes, M. J. Med. Chem. 1992,
35, 1156–1165.
13. (a) Chow, C. P.; Shea, K. J.; Sparks, S. M. Org. Lett. 2002,
4, 2637–2640; (b) Labaziewicz, H.; Howell, B. A.; Lindf-
ors, K. R.; Mason, R. C. Heterocycles 1992, 34, 699–711.
14. Perrault, W. R.; Pearlman, B. A.; Godrej, D. B.; Jegana-
than, A.; Yamagata, K.; Chen, J. J.; Lu, C. V.; Herrinton,
P. M.; Gadwood, R. C.; Chan, L.; Lyster, M. A.;
Maloney, M. T.; Moeslein, J. A.; Greene, M. L.; Barb-
achyn, M. R. Org. Process Res. Dev. 2003, 7, 533–546.
15. Urogdi, L.; Patthy, A.; Moravcsik, E.; Kisfaludy, L.,
et al. J. Heterocycl. Chem. 1989, 26, 129.
Oxazolidinones 6 and 20 were dosed orally in an experi-
mental murine in vivo infection model.23 Based on the
in vitro potency of these compounds in the presence of
50% calf serum, protein binding does not appear to ex-
plain the lack of in vivo activity. Since these oxazolidi-
nones had modest to poor activity (PD50 >50 mg/kg/
day) in this in vivo model, no further characterization
was undertaken.
In summary, oxazolidinones related to linezolid and
eperezolid were synthesized24 and found to exhibit
potent in vitro activity against a panel of Gram-positive
bacteria. The dihydro-1,2-oxazine present in 6 and the
2-pyrazoline present in 20 showed the most promise as
morpholine and piperazine replacements, respectively.
16. Riddel, F. G.; Turner, E. S.; Katritzky, A. R.; Patel, R. C.;
Brito-Palma, F. M. S. Tetrahedron 1979, 35, 1391.
17. Williams, D. R.; Osterhout, M. H. J. Am. Chem. Soc.
1992, 114, 8750–8751.
18. (a) Ihara Chem Ind Co Ltd; Japan Patent, JP-04244067,
92-337745, 1992; (b) Boros, E. E.; Bouvier, F.; Rabino-
witz, M. H. J. Heterocycl. Chem. 2001, 38, 613–
616.
Acknowledgements
The authors would like to thank the analytical chemistry
group at Bristol-Myers Squibb for support in character-
izing the compounds reported herein.
References and notes
19. The minimum inhibitory concentrations (MICs) were
determined using
a
broth microdilution assay in
1. Barbachyn, M. R.; Ford, C. W. Angew. Chem., Int. Ed.
2003, 42, 2010–2023.
accordance with that recommended by the National
Committee for Clinical Laboratory Standards (NCCLS).

S. DÕAndrea et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2834–2839
2839
Mueller-Hinton medium was used except for streptococci,
which was tested in Todd Hewitt broth. The final bacterial
inoculate contained approximately 5 · 105 cfu/mL and the
plates were incubated at 35 ꢁC for 18 h in ambient air
(streptococci in 5% CO2). The MIC was defined as the
lowest drug concentration that prevented physical growth.
20. Park, C. H.; Brittelli, D. R.; Wang, C. L. J.; Marsh, F. D.;
Gregory, W. A.; Wuonola, M. A.; McRipley, R. J.;
Eberly, V. S.; Slee, A. M.; Forbes, M. J. Med. Chem. 1992,
35, 1156–1165.
repeated. Also, compound 25 had MICÕs of 8 lg/mL
against two other strains (S. pneumoniae A27881 and S.
pneumoniae A28272) not included in Table 1.
22. Gregory, W. A.; Brittelli, D. R., et al. J. Med. Chem.
1990, 33, 2569–2578.
23. Staphylococcus aureus/Pen+ A15090, Inoculum 1.05 ·
107 cfu/mouse ip in 7% mucin, drugs dissolved in 10%
DMSO, 5% Tw-80, and water, administered orally b.i.d., 1
and 5 h post infection, death recorded for 8 days, 10 mice
per group.
21. The MIC of 25 against Streptococcus pneumoniae A9585 is
actually quite impressive, but this determination was not
24. All new compounds were characterized by analytical
methods (1H NMR and LC/MS).