Photochemistry of oxazolidinone antibacterial drugs

Article abstract of DOI:10.1111/j.1751-1097.2009.00546.x

The photochemistry of six N3-(3-fluoro-4-dialkylaminophenyl)-oxazolidinones known for their antimicrobial activity has been examined. All of these compounds are defluorinated in water (φ_dec ≈ 0.25) and in methanol (φ_dec ≈ 0.03), reas

Full text of DOI:10.1111/j.1751-1097.2009.00546.x

Photochemistry and Photobiology, 2009, 85: 879–885
Photochemistry of Oxazolidinone Antibacterial Drugs†
Elisa Fasani*, Fedele Tilocca and Angelo Albini
Department of Organic Chemistry, University of Pavia, Pavia, Italy
Received 23 December 2008, accepted 15 January 2009, DOI: 10.1111 ⁄ j.1751-1097.2009.00546.x
ment of thrombocytopenia upon prolonged treatment (9–11).
ABSTRACT
However, the presence of the fluoroaminophenyl motif is
reminiscent of the high photoreactivity of fluoroanilines (12)
and of the fact that other antimicrobials likewise with a
fluorine atom flanking an amino group such as lomefloxacin
and other fluoroquinolones are known for their high photo-
lability (13,14) and phototoxicity (15). Linezolid is known to
be liable to degradation under photochemical conditions (16–
18) and a recent detailed study has shown that the main
photoprocess does in fact involve the aromatic fluorine (19).
The reaction proceeded from the triplet state and was well
rationalized on the basis of a phenyl cation as the intermediate.
As mentioned above, the research on new oxazolidinones is
quite active at present (20–25) and several new derivatives are
candidates to the introduction in therapy. This encouraged us
to extend the photochemical study on some further derivatives
in order to understand whether structural variations had an
important effect on the photochemical properties. The choice
of the derivatives for the study was based on the significance of
the pharmacological properties.
The photochemistry of six N3-(3-fluoro-4-dialkylaminophenyl)-
oxazolidinones known for their antimicrobial activity has been
examined. All of these compounds are defluorinated in water
(Fdec ꢀ 0.25) and in methanol (Fdec ꢀ 0.03), reasonably via the
triplet. The chemical processes observed are reductive defluori-
nation and solvolysis, depending on the structural variation
introduced (thus, tethering the dialkylamino group to the
aromatic ring and introducing a highly polar group in the
oxazolidinone moiety have an effect). A likely mechanism
involves the fragmentation of the C–F bond yielding the
corresponding triplet phenyl cation. This intermediate either is
reduced or, under appropriate conditions, intersystem crosses to
the singlet state that adds the solvent. These data demonstrate a
sizeable photodecomposition of these drugs that causes a
decrease in the therapeutic activity. Furthermore, the likely
formation of phenyl cations may cause a photogenotoxic effect.
INTRODUCTION
In the last decades, the development of antibacterials has been
largely based on the improvement of existing classes, as the
introduction of a new family requires major investments.
Therefore, the emerging of oxazolidinones after a long interval
and after brilliant work by chemists and biologists is to be
considered a major advancement (1–5), culminating in the
licensing of the first clinically useful antibacterial of this class,
linezolid 1. In this molecule the heterocyclic moiety bears an
aliphatic side chain in position 5 (S configuration necessary for
drug activity) and a phenyl group on the nitrogen atom,
likewise required for activity. In turn, the phenyl moiety is
functionalized by fluorination (that increases the potency and
enhances the oral pharmacokinetic performance) and by
introducing an electron donating group (morpholino), well
tolerated and improving the safety profile. This drug is used
for the treatment of infections caused by multi-resistant
bacteria including streptococci (6) and methicillin-resistant
Staphylococcus aureus (7) and has many therapeutic indica-
tions, including the treatment of tuberculosis. Although this is
a rather expensive drug, treatments are considered more cost
effective than using competing drugs (8).
MATERIALS AND METHODS
H (300 MHz) and C (75.4 MHz) NMR spectra were registered by
means of a Brucker instrument and IR spectra by using a Perkin Elmer
Fourier transform spectrophotometer. Flash silica gel was used for
column chromatographic separation. Reverse-phase HPLC analysis
was carried out by using
a C18 PHENOMENEX Gemini,
25 cm · 4.6 mm, 5 lm column and eluting with H2O ⁄ MeCN 70 ⁄ 30
for 2, 6 and 7, 60 ⁄ 40 for 4 and 5, 50 ⁄ 50 for 3. The same setup was used
for HPLC ⁄ MS experiments. Fluorescence and phosphorescence spec-
tra were measured by means of a Perkin Elmer LS 55 luminescence
spectrometer under the conditions indicated in Table 4.
Starting materials. (S)-3-[3¢-Fluoro-4¢-(N-morpholino)phenyl]-
5-(N-acetamidomethyl)-oxazolidin-2-one (linezolid, 1) was prepared
according to the published procedure (5). N1-{(5S)-3-[(6aS)-1-Fluoro-
6a,7,8,9-tetrahydro-6H-azolo-[1,2-d]benzo[b][1,4]oxazin-3-yl]-2-oxo-
1,3-oxazolan-5-yl}methylacetamide 2 and N1-{(5S)-3-[(6aS)-6a,7,8,9-
tetrahydro-6H-azolo-[1,2-d]benzo[b][1,4]oxazin-3-yl]-2-oxo-1,3-oxazo-
lan-5-yl}methylacetamide 7 were prepared essentially according to the
published procedure (26) but because of low yields and difficult
application two of the intermediates involved were synthesized in a
different way as indicated below.
(6aS)-1-fluoro-6a,7,8,9-tetrahydro-6H-azolo-[1,2-d]benzo[b][1,4]
oxazine-3-amine: To a solution of 9.00 g (37.8 mmol) of (6aS)-1-
fluoro-3-nitro-6a,7,8,9-tetrahydro-6H-azolo-[1,2-d]benzo[b][1,4]oxazine
(26) in 50 mL of tetrahydrofuran and 90 mL of methanol, ammonium
formate was added (19.00 g). The flask was maintained under argon
atmosphere and cooled to 0ꢀC. Ten percent palladium on carbon
(0.44 g) was added and the suspension was stirred overnight. The
reaction was monitored by TLC, water and ethyl acetate were added,
the phases separated, and the aqueous portion was extracted with ethyl
acetate. The combined organic portions were washed with saturated
Linezolid is considered a relatively safe drug and the main
toxicity issue reported to date refers to the possible develop-
†This invited paper is part of
Photochemistry.
*Corresponding author email: elisa.fasani@unipv.it (Elisa Fasani)
a Symposium-in-Print on Pharmaceutical
ꢁ 2009 The Authors. JournalCompilation. The AmericanSocietyofPhotobiology 0031-8655/09
879

880 Elisa Fasani et al.
sodium chloride, dried (MgSO4) and evaporated to give 7.2 g of the
title compound (90%) as a brown solid that was sufficiently pure for
the following steps, through which it was transformed into (5R)-
3-[(6aS)-1-fluoro-6a-7,8,9-tetrahydro-6H-azolo-[1,2-d]benzo[b][1,4]
oxazine-3-yl]-5-azidomethyl-1,3-oxazolan-2-one according to the
reported method (26).
Compound 2: To a solution of 0.90 g (2.68 mmol) of the above azide
in 80 mL of ethyl acetate 10% palladium on carbon was added (0.25 g)
and the vessel was alternatively evacuated and filled with nitrogen;
then hydrogen was introduced. The mixture was stirred overnight, then
evacuated, flushed with nitrogen and cooled at 0ꢀC, when 0.25 mL
(3.08 mmol) of pyridine and 0.80 mL (8.3 mmol) of acetic anhydride
were added. The mixture was stirred for 30 min and then removed
from the ice bath and stirred at room temperature (r.t.) for 1 h, filtered
and concentrated to give an orange solid which was purified on a silica
gel column by chromatography to obtain 0.6 g (64%) of acetamide 2.
The corresponding intermediates for the synthesis of compound 7
were prepared by the same method.
identified on the basis of their analytical and spectroscopic properties.
The most informative evidence was offered by H, C NMR spectra as
reported below.
9
1H NMR (CDCl3): d 1.5–2.2 (m, 4H), 3.25 (q, J = 9 Hz, 1H),
3.45 (t, J = 10 Hz, 1H), 3.6 (m, 1H), 3.75–4.15 (m, 5H), 4.05 (s, 3H),
4.45 (dd, J = 10, 3 Hz, 1H), 4.95 (m, 1H), 6.50 (d, J = 8.5 Hz, 1H),
6.90–7.05 (m, 2H); 13C NMR (CDCl3): d 23.5 (CH2), 28.2 (CH2), 47.6
(CH2), 47.9 (CH2), 48.3 (CH2), 55.1 (CH), 57.5 (CH3), 68.6 (CH2), 70.1
(CH), 108.2 (CH), 112.2 (CH), 113.5 (CH), 127.0 (C), 132.6 (C), 142.8
(C), 154.6 (C = O), 192.5 (C = S).
1
10 H NMR (CDCl3): d 1.5–2.25 (m, 4H), 2.60 (q, J = 9 Hz, 1H),
3.80 (s, 3H), 4.05 (s, 3H), 3.30–4.10 (m, 8H), 4.85 (m, 1H), 6.45 (d,
J = 2.5 Hz, 1H), 6.95 (bs, 1H), 7.05 (d, J = 2.5 Hz, 1H); 13C NMR
(CDCl3): d 23.0 (CH2), 26.0 (CH2), 47.6 (CH2), 47.8 (CH2), 51.8 (CH2),
54.1 (CH), 55.7 (CH3), 57.6 (CH3), 64.8 (CH2), 71 (CH), 95.5 (CH),
100.0 (CH), 120.7 (C), 131.6 (C), 146.5 (C), 153.1 (C), 154.6 (C = O),
192.5 (C = S).
16 1H NMR (CDCl3): d 3.10 (t, J = 4.5 Hz, 4H), 3.85 (t,
J = 4.5 Hz, 4H), 4.0 (s, 3H), 3.75–4.10 (m, 4H), 4.95 (m, 1H), 6.90
(d, J = 9 Hz, 2H), 7.25 (bs, 1H), 7.40 (d, J = 9 Hz, 2H); 13C NMR
(CDCl3): d 47.50 (CH2), 47.60 (2 CH2), 48.10 (CH2), 57.50 (CH3),
66.70 (2 CH2), 71.10 (CH), 116.0 (2 CH), 120.0 (2 CH), 130.3 (C),
148.2 (C), 154.6 (C = O), 192.6 (C = S).
N1-{(5S)-3-[(6aS)-1-Fluoro-6a,7,8,9-tetrahydro-6H-azolo-[1,2-d]
benzo[b][1,4]oxazin-3-yl]-2-oxo-1,3-oxazolan-5-yl}methyl-O-methylthio-
carbamate 3 and N1-{(5S)-3-[(6aS)-1-Fluoro-6a,7,8,9-tetrahydro-6H-
azolo-[1,2-d]benzo[b][1,4]oxazin-3-yl]-2-oxo-1,3-oxazolan-5-yl}methyl-
thiourea 4 were prepared according to the published procedure (26).
(S)-N-{[3-(3-fluoro-4-morpholinophenyl)-2-oxo-1,3-oxazolan-5-yl]
17 1H NMR [(CD3)2CO]: d 2.85 (t, J = 5 Hz, 4H), 3.90 (s, 3H),
3.80 (t, J = 5 Hz, 4H), 3.80–4.10 (m, 4H), 4.95 (m, 1H), 7.0 (dd,
J = 9, 2.5 Hz, 1H), 7.25 (d, J = 9 Hz, 1H), 7.3 (d, J = 2.5 Hz, 1H),
7.8 (bs, 1H), 8.5 (bs 1H); 13C NMR [(CD3)2CO]: d 48.9 (CH2), 49.1
(CH2), 53.6 (2 CH2), 57.8 (CH3), 67.8 (CH2), 71.7 (CH), 106.5 (CH),
110.7 (CH), 122.0 (CH), 136.8 (C), 137.5 (C), 152.9 (C = O), 193.9
(C = S).
Quantum yield measurements. Reaction quantum yields were mea-
sured by irradiating 2 mL samples of 5 · 10)4 M solutions of 1–7 in a
quartz spectrophotometric cuvette on an optical bench. The light
source was a collimated beam from a 100 W high-pressure mercury arc
fitted with an interference filter (transmittance maximum, 280 nm).
The reaction was monitored by HPLC and the consumption of the
starting material (limited to <20%) was determined (a known volume,
20 lL, injection loop was used). The light flux was measured by
ferrioxalate actinometry (29).
methyl}-O-methylthiocarbammate
5
morpholinophenyl)-2-oxo-1,3-oxazolan-5-yl]methyl}thiourea
and (S)-N-{[3-(3-fluoro-4-
were
6
prepared analogously to 3 and 4 from (S)-5-Aminomethyl-[N-3-(3-
fluoro-4-morpholinophenyl)]oxazolidine-2-one as indicated below.
(R)-{[N-3-(3-fluoro-4-morpholinophenyl)-2-oxo-oxazolidin-5-yl}
methyl]isothiocyanate: Thiophosgene (1.25 mL) was added dropwise
to a solution of (S)-5-Aminomethyl-N-[3-(3-fluoro-4-morpholinophe-
nyl)]oxazolidine-2-one (4.7 g, 14.13 mmol) (27) in dry dichlorome-
thane (100 mL) in an ice bath under argon. The reaction mixture was
allowed to react at r.t. over 3 h and then the volatile components were
removed. The residue obtained was chromatographed onto a silica gel
column and afforded 2 g of the title compound (40%). 1H NMR
(CDCl3) d 3.15 (t, J = 4.5 Hz, 4H), 3.90 (t, J = 4.5 Hz, 4H), 3.80–
4.04 (m, 3H), 4.15 (t, J = 9 Hz, 1H), 4.85 (m, 1H), 6.95 (t, J = 9 Hz,
1H), 7.15 (dd, J = 9, 2.5 Hz, 1H), 7.45 (dd, J = 14–2.5 Hz, 1H).
Compound 5: A solution of the above isothiocyanate (1.7 g,
5 mmol) in methanol (100 mL) was heated to 80–100ꢀC while
monitoring by TLC. When the consumption of the starting material
was complete, the reaction mixture was allowed to cool to r.t. The
residue was then filtered, washed with ether and dried to afford a crude
which was crystallized from isopropanol ⁄ water to give 1.3 g of 5
(70%), the spectroscopic properties of which corresponded to those
reported in the literature (27).
Fluorescence quantum yields were measured using quinine sulfate
as the standard (FF = 0.54) (30).
RESULTS
First reported by Dupont researchers in 1987 (3) oxazolidi-
none antimicrobials have been the subject of extensive inves-
tigation and development at the Upjohn (then Pharmacia)
laboratories (4,5). A member of this family, linezolid, was
approved by the Food and Drug Administration in April 2000
and has acquired a significant therapeutic role. In the
following years, the success of this drug fostered structure–
activity relationship (SAR) studies for the optimization of the
pharmaceutical activity of molecules having as the base
skeleton a N-(aminophenyl)-oxazolidinone and bearing a polar
group tethered through a methylene in position 5. These
studies showed that important elements are the introduction of
a conformational constraint in the aminophenyl moiety, the
introduction of a fluoro substituent on the phenyl ring and
the nature of the polar group (31). Due to our interest in the
photochemistry of drugs and in particular of those containing
a fluorinated aromatic moiety, we decided to examine the
photochemistry of some of these compounds, as a contribution
to the pharmacological profile of such highly useful drugs.
Thus, we compared the photochemistry of some fluor-
ophenyloxazolidinones that appeared promising for the
biological activity. In these the structural modifications that
the above SAR analysis had indicated as significant were
introduced. Thus, the polarity of the side chain was varied, the
Compound 6: Ammonia gas was bubbled into a solution of the
above isothiocyanate (300 mg, 0.89 mmol) in THF (100 mL) at )10ꢀC
over 20 min. The resultant mixture was stirred at r. t. for 1 h and then
diluted with ethyl acetate. The organic layer was extracted with water
and brine and dried. The solvent was evaporated and the residue
obtained was passed through a column of silica gel to afford 170 mg of
thiourea (6) (54%), the spectroscopic properties of which corre-
sponded to those reported in the literature (28).
Photochemical reactions. Explorative experiments: 5 · 10)4 M Solu-
tions of the investigated compounds were flushed by nitrogen and
irradiated by means of six external phosphor-coated lamps (15 W,
center of emission, 310 nm), monitoring the course of the reaction by
HPLC. Under these conditions, the light flux measured by
a
radiometer fitted by a UV-31 calibrated sensor was 61 W m)2. Tests
were carried out also by using lamps with center of emission at 360 nm
in the same setup (flux measured by a radiometer fitted by a UV-36
calibrated sensor was 50 W m)2).
Preparative experiments: 1 · 10)3 M Solutions of the investigated
compounds were flushed by nitrogen and irradiated under either of the
two conditions, viz. (1) in an immersion well apparatus (125 mL) by
means of a 125 W medium pressure mercury arc through Pyrex or (2)
in a number of quartz tubes (each containing 10 mL) by means of four
external phosphor-coated lamps (15 W, center of emission, 310 nm).
The course of the reaction was monitored by HPLC. When the starting
material was consumed, the solvent was removed under reduced
pressure and the crude product was purified by flash chromatography
on silica gel (cyclohexane-ethyl acetate as eluent). New products were

Photochemistry and Photobiology, 2009, 85 881
Table 1. Products from the irradiation of tricyclic oxazolidinones 2–4.
rotation of the dialkylaminophenyl group was hindered and
fluorinated and fluorine-free derivatives were compared. The
first group of oxazolidinones included tricyclic compounds 2, 3
and 4, where an oxymethylene bridge hinders the rotation of
the morpholino group with respect to the aromatic ring; these
compounds bear an acetamido, a thiocarbamate or a thiourea
polar group, respectively. The second group included the
corresponding derivatives lacking the oxymethylene bridge,
viz. compounds 1 (the previously examined linezolid), 5 and 6
(Scheme 2). The photochemistry was systematically investi-
gated in water and in methanol by irradiating at 310 nm; tests
with compounds 1 and 2 showed a significant photodecom-
position also by using lamps centered at 360 nm (ca 3% of the
decomposition observed with the 310 nm lamps). Exposure of
a similar aqueous solution of 1 and 2 to an integrated near-UV
energy of 200 W m)2 (360 nm centered lamps) caused a
decomposition, respectively, of 56% and 65% for the two
drugs. For some derivatives the photoreactivity was tested also
in nonpolar solvents and found to be very low; therefore, such
conditions were not further investigated.
N1-{(5S)-3-[(6aS)-1-Fluoro-6a,7,8,9-tetrahydro-6H-azolo-
[1,2-d]benzo[b][1,4]oxazin-3-yl]-2-oxo-1,3-oxazolan-5-yl}meth-
ylacetamide 2 was prepared through a modification of the
published procedure (see Materials and Methods). Irradiation
(phosphor-coated lamps, emission centered at 310 nm) of this
compound (1 · 10)3 M) in a nitrogen-flushed aqueous solution
led to extensive degradation (85% in 15 min), as indicated by
HPLC. A single main product was formed and was separated
by column chromatography (a few highly polar peaks were
revealed, but not separated). This was demonstrated to be the
corresponding fluorine-free phenyloxazolidinone 7 by com-
parison with an authentic sample prepared according to the
literature (26) (see Scheme 1 and Table 1).
Oxazolin
Solvent
Products (% yield)
2
3
4
H2O
MeOH
H2O
MeOH
H2O
MeOH
7 (48)
7 (35), 8 (18)
9 (34)
9 (28), 10 (13)
11 (35)
11 (25), 12 (25)
the only isolated product in water and as the main one in
methanol, where methyl ether 10 was also formed. As for
thiourea 4, the HPLC profile of the irradiated solution in both
water and methanol was almost identical to that obtained from
3 under the same conditions and it was assumed that the
photoproducts were 11 and 12, the analogs of those from 3.
The nonfluorinated derivative 7 was irradiated under the
same conditions and showed to be next to photostable. A
single peak was detected in HPLC after prolonged irradiation.
Examination by mass spectroscopy demonstrated the loss of 2
mass units with respect to the starting material and the
compound was suggested to have structure 13.
The photochemistry of N-[3-(3-fluoro-4-morpholinophe-
nyl)-2-oxo-oxazolidin-5-yl]methylacetamide (1, linezolid) has,
as mentioned, been examined previously (19) and found to give
the defluorinated product 14 in 30% yield in water. For some
further derivatives, like fluorine-free 14, it had been demon-
strated that the morpholine ring had undergone a slow
oxidative degradation (structure 15 and three products derived
from it). Minor, highly polar products were detected, but not
further investigated. Reduced 14 and a trace of the methyl
ether were obtained in methanol. For the convenience of the
reader, the main results are reported again in Scheme 2a and
Table 2.
When the same compound was irradiated in methanol,
reduced 7 was again the main photoproduct and was accom-
panied by a lower amount of another phenyloxazolidinone
that was separated. The evidence obtained allowed to assign
the structure of methyl ether 8 to this compound.
Two further oxazolidinones, 3 and 4, were prepared
through modifications of the published procedure (see Mate-
rials and methods). These differed from compound 2 by having
a different group attached to the aminomethyl chain in
position 4, and contained a methylthiourethane and a thioure-
ide function. Irradiation of compound 3 gave fluorine-free 9 as
Two analogs of compound 1 with a different polar group in
position 5 were prepared by conventional syntheses (see
Materials and Methods), viz. the methyl urethane 5 and the
ureide 6. Irradiation of compound 5 in water for 2 h led to a
45% decomposition. Two significant products were detected
by HPLC and were separated by column chromatography and
characterized. Analytical and spectroscopic examination
showed that the minor one had undergone reductive defluo-
rination (structure 16) and the main one substitution of a
Scheme 1.

882 Elisa Fasani et al.
Scheme 2.
Table 2. Products from the irradiation of oxazolidinones 1, 5 and 6.
Products
(% yield)
Oxazolin
Solvent
1
5
6
H2O
MeOH
H2O
MeOH
H2O
MeOH
14 (30)*
14 (70)†
16 (4), 17 (25)
16 (33)
18 (4), 19 (25)
18 (33)
*Products resulting from the oxidative degradation of the morpholino
side chain, viz. enamine 15 and products resulting from it, were also
obtained (ca 20%). †A 10% yield of the 3-methoxy-4-morpholino
derivative was also obtained (19).
Figure 1. Phosphorescence spectra of compounds 1, 2, 7 and 14 in
ether-pentane-alcohol (5 ⁄ 5 ⁄ 2) at 77 K.
Table 3. Quantum yield of reaction of oxazolidinones 1–7.*
Table 4. Key spectroscopic data on some oxazolidinones.
Oxazolin
H2O
MeOH
Absorbance*
Fluorescence*
Phosphorescence†
2
3
4
1
5
6
7
0.27
0.09
0.11
0.33
0.25
0.23
<0.001
0.02
0.02
0.03
0.05
0.05
0.04
0.03
Relative
intensity‡
k
max, nm (log ꢀ)
k
max, nm
FF‡
k
max, nm
2
7
1
14
3
247 (4.33)
264 (4.04)
252 (4.29)
255 (4.29)
259 (4.16)
246 (4.22)
377§
358||
377§
358§
0.03
0.11
0.0004
0.07
427§
454||
445§
427§
0.23
0.9
0.02
1
*Error within 10%.
5
hydroxyl group for a fluoro atom (structure 17). In methanol
compound 16 was the only product. As for ureide 7, this
showed a HPLC strictly analogous to that of 5 in both solvents
and was assumed to follow the same course giving 18 and 19
(Scheme 2b, Table 2).
*In water at 20ꢀC. †In ether-pentane-ethyl alcohol 5:5:2 mixed solvents
at 77 K. ‡Error within 10%. §kexc 265 nm; slit bandwidths (nm) 10–10.
||kexc 290 nm; slit bandwidths (nm) 10–10.
The quantum yield of reaction of the above compounds was
measured in the solvents used for preparative reactions (see
Table 3).
Some spectroscopic properties of representative compounds
were measured. The absorbance spectra were quite similar,
with a maximum at ca 250–270 nm. The data for tricyclic
derivatives 2 and 3, compounds 1 and 5 lacking the oxygenated
ring and the related fluorine-free compounds 7 and 14 are
reported in Table 4.
The emission properties of some of the above oxazolidi-
nones were also examined. The fluorophenyl derivatives 1 and
2 were only weakly fluorescent in solution, while the fluorine-
free analogs were 30 times more emissive. Likewise, fluorinated
2 phosphoresced weakly in glass at 77 K and the emission

Photochemistry and Photobiology, 2009, 85 883
from 1 was barely detectable, while the phosphorescence from
fluorine-free 7 and 14 was much more intense (see Fig. 1 and
Table 4).
the photoreactions considered, with quantum yield up to 0.33,
originates from the triplet (320 in Scheme 3). The type of
chemistry observed then supports this idea, as shown below.
Two mechanisms appear reasonable (19), that is monomo-
lecular cleavage of the carbon–fluorine bond (SN1 process,
path a in Scheme 3) and bimolecular [SN2 (Ar*) path b]
reaction via the addition–elimination mechanism usual in the
thermal chemistry of aromatics. Note that in a recent flash
photolysis study on linezolid 1 (19), we documented a further
process, photoionization. This was proved to be a reversible
process and to cause no irreversible chemical change. This can
thus be disregarded in the present chemical investigation.
In the SN1 case, homolytic cleavage of the C–F bond is
excluded because the energy of both singlet (ES around
85 kcal mol)1) and triplet (ET around 75 kcal mol)1) states,
as measured from the emission spectra, are too low to make
this a viable path (EAr-F ꢀ 120 kcal mol)1). Thus, a mono-
molecular mechanism must involve heterolytic cleavage to
form a phenyl cation. The two mechanisms are depicted in
Scheme 3.
DISCUSSION
The absorption spectra of the oxazolidinones considered in
this study show little difference among them (see Table 4).
Oxazolidinones bearing no aromatic substituent do not absorb
above 250 nm (32). When an aromatic substituent is present,
as in the compounds above, the spectrum is essentially that of
that moiety, with little if any modification by the oxazolidi-
none moiety. In fact, the spectra of the present compounds all
are closely analogous to that of anilines. Likewise, the
oxazolidinone moiety does not introduce any photochemical
reactivity. N-phenyl derivatives such as compounds 7 and 14
are rather photostable and actually the only reaction occurring
involves the morpholino group substituted at the phenyl ring
(the usual oxidative side chain degradation previously ob-
served in N-dialkylanilines), not the oxazolidinone moiety.
Furthermore, such fluorine-free compounds exhibit both
fluorescence in solution and phosphorescence in glass (see
Table 4) and resemble anilines in their photophysical proper-
ties. Comparison with literature data shows that the fluores-
cence of parent aniline is at 334 nm (FF = 0.15 in MeCN) and
the phosphorescence at 405 nm (in methyltetrahydrofuran
glass at 77 K) (33). Thus, the emission from the present
derivatives is somewhat redshifted and less intense, but
qualitatively of the same type.
On the other hand, all of the fluorinated compounds tested
are quite photolabile and the photoreaction consistently
involves the fluoro atom through either of the two processes,
reductive elimination or solvolysis. In all of the cases, the
efficiency of the reaction strongly decreases in going from
water to methanol as the solvent and decreases again in passing
to acetonitrile. Thus it is likely that the reaction involves
charged intermediates. Insertion of a fluorine decreases both
the fluorescence and the phosphorescence of these compounds,
thus no indication about the multiplicity of the reacting states
comes from this fact. However, with simple anilines, including
some fluorinated anilines (34), intersystem crossing in solution
is quite efficient (FISC ꢀ 0.7) and thus it seems possible that
The two reactions observed, reduction and substitution,
may result each from one of the mechanisms (path a or b) or
both from path a, with a later branching through path c
toward substitution. Several studies on halogenated anilines
(35) have shown that reductive dehalogenation (including
3
defluorination) via the triplet phenyl cation (here 21) formed
from the cleavage of the triplet state is a viable path and it
appears reasonable that this applies also to this case. Triplet
phenyl cations do not add to n nucleophiles but rather abstract
hydrogens. Ongoing computational studies on analogous
reactions suggest that the triplet multiplicity is conserved
during the bond formation sequence and is followed by
intersystem crossing to the singlet end product. As for
solvolysis this may arise either again from the cation, in this
case the singlet (121, known to be the most stable spin state of
cations of this type) (36), provided that intersystem crossing
(path c) occurs efficiently, or via the addition–elimination
mechanism (path a). It is difficult to assess the two possibilities,
each of which has precedents for analogous compounds (37).
However, some trends can be evidenced. Thus, the quantum
yield of reaction is rather constant along the series, around
0.25 in water (except for compound 3 that has ca 0.1) and
Scheme 3.

884 Elisa Fasani et al.
0.02–0.05 in methanol, with a decrease by a factor of 5–10 in
going to the latter solvent. The limited variation along the
series suggests the primary photochemical step remains the
same. The effect of structure can be evidenced in the type of
chemical process, though. Thus, conformationally blocked
derivatives 2–4 undergo no substitution in water, but only (as a
minor process) in methanol, a fact that strongly supports the
phenyl cation path. Substitution on the contrary is the main
path in conformationally free compounds 5 and 6 in water, but
not with 1, and with these compounds no substitution takes
place in methanol (again an exception is 1, where this process
occurs, though to a low extent, see Table 2). As the two former
compounds are distinguished from the last one by a more
polar side chain in 5 (the dipole moment of N-methylacetamide
is 3.98 D in dioxane, while that of N,N’-dimethylthiourea is
5.31 D), an economic rationalization is that the photoreaction
proceeds in any case via heterolytic cleavage in the triplet state
(hence the much larger Fr in more polar water [e = 80.2] than
in MeOH [e = 32.6]) to give the triplet cation. An optimal
solvation of such strongly polar intermediate prolongs its
lifetime. This is the case when the cationic site is easily
accessible (that is in conformationally free compounds) and in
the presence of highly polar groups (that is the case of
compounds 5 and 6). Under these conditions ISC to the singlet
phenyl cation becomes competitive with reaction from the
triplet, particularly in a nonreducing solvent such as water,
and this is known to add to n nucleophiles. The polarity of the
side chain has little importance with compounds 2–4, where
the determining factor seems to be sterical hindering around
the reacting center. Apparently, the solvation in this case is
such that solvolysis is competitive only in methanol.
The relatively high quantum yield of reactions of these
molecules in water should be highlighted. Linezolid is
approved for the use also by parenteral infusion and in
particular under these conditions it should be protected from
light in order to avoid a serious loss of activity. In fact the
scission of F atom, that increases the potency of the drug,
should affect the antibacterial activity. As shown above, the
photodecomposition is significant also under ICH confirma-
tory test conditions (200 W h m)2). Thus, one cannot count
too much on the poor absorption of near-UV light by drug
preparations with compound 1 as the active principle. A
photoinduced lessening of the therapeutic activity of linezolid
and oxazolidinones of similar structure that may in the future
be used as drugs may result even after assumption. The
photoreaction may be linked with some phototoxic effect,
although this has not been reported as yet for the case of
compound 1, the only derivative presently used. The above
chemical data suggest that a phenyl cation is formed, just as in
the case of fluoroquinolones. In particular, the effective
trapping with pyrrole strengthens the idea that a similar
electrophilic intermediate is formed and attacks selectively
electron-rich heterocycles and, by analogy, DNA bases. This
points the attention (38) to the possible phototoxicity of the
oxazolidinones and, on the other hand, photoactivated drugs
based on this structure may be conceived.
reaction involves the moiety present in such molecules,
probably via heterolysis of the C–F bond in the triplet state.
The chemical paths from the thus formed triplet phenyl
cation are then determined by the solvation and thus by the
rigidity and the accessibility of the reacting center as well as
by the presence of highly polar groups in the molecule.
Although these compounds do not absorb significantly above
310 nm, exposure to UV-B light can cause
a relevant
decomposition and thus a loss of therapeutic activity. As
the reaction leads to phenyl cations, suspected to have an
important role in the phototoxicity of other drugs, in
particular fluoroquinolones, the photobiological implications
of these results should be taken into account and foster an
appropriate investigation.
REFERENCES
1. Barbachyn, M. R. and C. W. Ford (2003) Oxazolidinone struc-
ture–activity relationships leading to linezolid. Angew. Chem.
Int. Ed. Engl. 42, 2010–2023.
2. Barrett, J. F. (2000) Linezolid Pharmacia Corp. Curr. Opin.
Investig. Drugs 1(2), 181–187.
3. Slee, A. M., M. A. Wuonola, R. J. McRipley, I. Zajac,
M. J. Zawada, P. T. Bartholomew, W. A. Gregory and M. Forbes
(1987) Oxazolidinones, a new class of synthetic antibacterial
agents: In vitro and in vivo activities of DuP 105 and DuP 721.
Antimicrob. Agents Chemother. 31(11), 1791–1797.
4. Barbachyn, M. R., S. J. Brickner, G. J. Cleek, R. C. Gadwood, K.
C. Grega, S. K. Hendges, D. K. Hutchinson, P. R. Manninen and
K. Munesada (1997) Design and synthesis of novel oxazolidinones
active against multidrug-resistant bacteria. Spec. Publ.—RSC 198,
15–26.
5. Brickner, S. J., D. K. Hutchinson, M. R. Barbachyn,
P. R. Manninen, D. A. Ulanowicz, S. A. Garmon, K. C. Grega,
S. K. Hendges, D. S. Toops, C. W. Ford and G. E. Zurenko (1996)
Synthesis and antibacterial activity of U-100592 and U-100766,
two oxazolidinone antibacterial agents for the potential treatment
of multidrug-resistant gram-positive bacterial infections. J. Med.
Chem. 39(3), 673–679.
6. Burkhardt, O., M. W. Pletz, C. P. Mertgen and T. Welte (2007)
Linezolid—The first oxazolidinone in the treatment of nosocomial
MRSA pneumonia. Recent Patents Anti-Infect. Drug Disc. 2(2),
123–130.
7. Von Der Lippea, B., P. Sandvenb and O. Brubakk (2006) Efficacy
and safety of linezolid in multidrug resistant tuberculosis (MDR-
TB)—A report of ten cases. J. Infect. 52(2), 92–96.
8. Grau, S. and C. Rubio-Terres (2008) Pharmacoeconomics of
´
linezolid. Expert Opin. Pharmacother. 9(6), 987–1000.
9. Johnson, A. P. (2003) The future prospects of oxazolidinones.
Investing. Drugs J. 6(3), 240–245.
10. Hutchinson, D. K. (2003) Oxazolidinone antibacterial agents:
A critical review. Curr. Top. Med. Chem. 3(9), 1021–1042.
11. Bernard, L., R. Stern, D. Lew and P. Hoffmeyer (2003) Serotonin
syndrome after concomitant treatment with linezolid and cita-
lopram. Clin. Infect. Dis. 36(9), 1197.
12. Freccero, M., M. Fagnoni and A. Albini (2003) Homolytic vs
heterolytic paths in the photochemistry of haloanilines. J. Am.
Chem. Soc. 125(43), 13182–13190.
13. Monti, S., S. Sortino, E. Fasani and A. Albini (2001) Multifaceted
photoreactivity of 6-fluoro-7-aminoquinolones from the lowest
excited states in aqueous media: A study by nanosecond and
picosecond spectroscopic techniques. Chem. Eur. J. 7, 2185–2196.
14. Albini, A. and S. Monti (2003) Photophysics and photochemistry
of fluoroquinolones. Chem. Soc. Rev. 32(4), 238–250.
15. Marrot, L., J. Belaıdi, C. Jones, P. Perez and L. Riou (2003)
¨
Molecular responses to photogenotoxic stress induced by the
antibiotic lomefloxacin in human skin cells: From DNA damage
to apoptosis. J. Invest. Dermatol. 121, 596–606.
CONCLUSION
16. Martin, G. E., R. H. Robins, P. B. Bowman, W. K. Duholke,
K. A. Farley, B. D. Kaluzny, J. E. Guido, S. M. Sims,
Oxazolidinone antibacterials appear to be quite photoreactive
in water and in polar organic solvents. The (rather efficient)

Photochemistry and Photobiology, 2009, 85 885
T. J. Thamann, B. J. Thompson, T. Nishimura, Y. Noro and T.
Tahara (1999) Susceptibility of morpholine substituents to photo-
oxidative decomposition—Identification of photo-oxidative de-
gradants of linezolid. J. Heteroc. Chem. 36(1), 265–270.
17. Agrawal, H., K. R. Mahadik and N. Paradkar (2003) Stability
indicating HPTLC determination of linezolid as bulk drug and in
pharmaceutical dosage form. Drug Dev. Ind. Pharm. 29(10), 1119–
1126.
18. Rao, B. M., M. K. Srinivasu, G. Sridhar, P. R. Kumar, K. B.
Chandrasekhar and A. Islam (2005) A stability indicating LC
method for zolmitriptan. J. Pharm. Biomed. Anal. 39, 503–
509.
19. Fasani, E., F. Tilocca, S. Protti, D. Merli and A. Albini (2008)
Photoionization and photoheterolytic defluorination in the drug
linezolid and related molecules. SN1(Ar*) vs. SR+N1(Ar*)
mechanism. Org. Biomol. Chem. 6, 4634–4642.
20. Renslo, A. R., P. Jaishankar, R. Venkatachalam, C. Hackbarth, S.
Lopez, D. V. Patel and M. F. Gordeev (2005) Conformational
constraint in oxazolidinone antibacterials. Synthesis and struc-
ture–activity studies of (azabicyclo[3.1.0]hexylphenyl)oxazolidi-
nones. J. Med. Chem. 48(15), 5009–5024.
21. Selvakumar, N., B. Yadi Reddy, G. Sunil Kumar, M. K. Khera,
D. Srinivas, M. Sitaram Kumar, J. Das, J. Iqbal and S. Trehan
(2006) Synthesis of novel tricyclic oxazolidinones by a tandem
SN2 and SNAr reaction: SAR studies on conformationally con-
strained analogues of linezolid. Bioorg. Med. Chem. Lett. 16(16),
4416–4419.
22. Srivastava, B. K., M. R. Jain, M. Solanki, R. Soni, D. Valani, S.
Gupta, B. Mishra, V. Takale, P. Kapadnis, H. Patel, P. Pandya, J.
Z. Patel and P. R. Patel (2008) Synthesis and in vitro antibacterial
activities of novel oxazolidinones. Eur. J. Med. Chem. 43(4), 683–
693.
23. Vara Prasad, J. V. (2007) New oxazolidinones. Curr. Opin.
Microbiol. 10(5), 454–460.
24. Komine, T., A. Kojima, Y. Asahina, T. Saito, H. Takano, T.
Shibue and Y. Fukuda (2008) Synthesis and structure–activity
relationship studies of highly potent novel oxazolidinone anti-
bacterials. J. Med. Chem. 51(20), 6558–6562.
activity relationship (SAR) studies on selected novel tricyclic
oxazolidinones. J. Med. Chem. 45(18), 3953–3962.
27. Tokuyama, R., Y. Takahashi, Y. Tomita, M. Tsubouchi, N.
Iwasaki, N. Kado, E. Okezaki and O. Nagata (2001) Structure–
activity relationship (SAR) studies on oxazolidinone antibacterial
agents. 3. Synthesis and evaluation of 5-thiocarbamate oxazoli-
dinones. Chem. Pharm. Bull. 49(4), 361–367.
28. Tokuyama, R., Y. Takahashi, Y. Tomita, M. Tsubouchi, T.
Yoshida, N. Iwasaki, N. Kado, E. Okezaki and O. Nagata (2001)
Structure–activity relationship (SAR) studies on oxazolidinone
antibacterial agents. 2. Relationship between lipophilicity and
antibacterial activity in 5-thiocarbonyl oxazolidinones. Chem.
Pharm. Bull. 49(4), 353–360.
29. Hatchard, C. G. and C. A. Parker (1956) A new sensitive chemical
actinometer. II. Potassium ferrioxalate as a standard chemical
actinometer. Proc. R. Soc. London, Ser. A 235, 518–536.
30. Dawson, W. R. and M. W. Windsor (1968) Fluorescence yields of
aromatic compounds. J. Phys. Chem. 72(9), 3251–3260.
31. Barbachyn, M. R., G. J. Cleek, L. A. Dolak, S. A. Garmon, J.
Morris, E. P. Seest, R. C. Thomas, D. S. Toops, W. Watt, D. G.
Wishka, C. W. Ford, G. E. Zurenko, J. C. Hamel, R. D. Schaadt,
D. Stapert, B. H. Yagi, W. J. Adams, J. M. Friis, J. G. Slatter, J.
P. Sams, N. L. Oien, M. J. Zaya, L. C. Wienkers and M. A.
Wynalda (2003) Identification of phenylisoxazolines as novel and
viable antibacterial agents active against Gram-positive patho-
gens. J. Med. Chem. 46(2), 284–302.
32. Schultze, H. (1973) Photolysis of N-substituted O-alkylurethanes.
Z. Naturforsch. (B) 28(5–6), 339–352.
33. Lewis, F. D. and W. Liu (1999) Luminescence of N-arylbenza-
mides in low-temperature glasses. J. Phys. Chem. A 103(48), 9678–
9686.
34. Othmen, K., P. Boule, B. Szczepanik, K. Rotkiewicz and G.
Grabner (2000) Photochemistry of 4-chloroaniline in solution.
Formation and kinetic properties of a new carbene, 4-iminocy-
clohexa-2,5-dienylidene. J. Phys. Chem. A 104(42), 9525–9534.
35. Guizzardi, B., M. Mella, M. Fagnoni, M. Freccero and A. Albini
(2001) Generation and reactivity of the 4-aminophenyl cation by
photolysis of 4-chloroaniline. J. Org. Chem. 66(19), 6353–6363.
36. Lazzaroni, S., D. Dondi, M. Fagnoni and A. Albini (2008)
Geometry and energy of substituted phenyl cations. J. Org. Chem.
73(1), 206–211.
25. Selvakumar, N., G. G. Rajulu, K. C. Reddy, B. C. Chary, P. K.
Kumar, T. Madhavi, K. Praveena, K. H. Reddy, M. Takhi, A.
Mallick, P. V. Amarnath, S. Kandepu and J. Iqbal (2008) Syn-
thesis, SAR, and antibacterial activity of novel oxazolidinone
analogues possessing urea functionality. Bioorg. Med. Chem. Lett.
18(2), 856–860.
26. Selvakumar, N., D. Srinivas, M. K. Khera, M. S. Kumar, R. N.
Mamidi, H. Sarnaik, C. Charavaryamath, B. S. Rao, M. A.
Raheem, J. Das, J. Iqbal and R. Rajagopalan (2002) Synthesis of
conformationally constrained analogues of linezolid: Structure–
37. Fasani, E., F. F. Barberis Negra, M. Mella, S. Monti and A.
Albini (1999) Photoinduced C–F bond cleavage in some fluori-
nated 7-amino-4-quinolone-3-carboxylic acids. J. Org. Chem.
64(15), 5388–5395.
38. Martinez, L. and C. F. Chignell (1998) Photocleavage of DNA by
the fluoroquinolone antibacterials. J. Photochem. Photobiol. B,
Biol. 45, 51–59.