An improved industrial synthesis of florfenicol plus an enantioselective total synthesis of thiamphenicol and florfenicol

Full text of DOI:10.1021/jo961479+

2996
J . Org. Chem. 1997, 62, 2996-2998
chloromethyl)oxazoline 4a and a regioselective epoxide
opening with CHCl2CN.
An Im p r oved In d u str ia l Syn th esis of
F lor fen icol p lu s a n En a n tioselective Tota l
Syn th esis of Th ia m p h en icol a n d
F lor fen icol
Resu lts a n d Discu ssion
The use of benzonitrile as a protecting group in our
previous report5b results in the inelegance of adding and
removing an amide moiety only to later replace it with
the desired one. Clearly, it would be shorter if dichlo-
roacetonitrile could be used instead of benzonitrile to
protect both the 1′-hydroxy and 2′-amino groups since the
protecting group ends up as a part of the product. As
shown in Scheme 1, we found that acids such as H2SO4
catalyze oxazoline formation at 70 °C. The resulting
(NH4)2SO4, in turn, promotes the isomerization of 4 to
the desired 5 at 50-60 °C. These two stages of reaction
temperature are necessary since the reaction takes place
in two phases. The first is the actual condensation and
cyclization at higher temperature to form a mixture of
the two oxazolines. The second is the precipitation-
driven equilibration of 4 to the desired internal oxazoline
5 at lower temperature. Filtration of the precipitate
improves the ratio of 5:4 from 97:3 to 99.5:0.5. Following
the fluorodehydroxylation procedure,7 5 was converted
cleanly to the fluorinated oxazoline 6. A selective hy-
drolysis procedure was developed to convert 6 to 2 in 86%
yield. This process has been scaled up successfully to
produce commercial florfenicol.
Having established 5 as a viable intermediate, we
turned our attention to its enantioselective synthesis. As
outlined in Scheme 2, the synthetic strategy is to regio-
selectively open the trans epoxide 7 with CHCl2CN,
forming 3′-(S,R)-(dichloromethyl)oxazoline 4a . Our ini-
tial attempt to form the imidate under acidic conditions
did not produce any desired product.8 Under basic
conditions, deprotonation of epoxy alcohol 7 with NaH
followed by addition of CHCl2CN did yield some imidate
which resulted in a few percent of 4a upon acidification.9
But the reaction mixture was complex. We suspected
that two major side reactions were hindering the desired
reaction. One is the simple deprotonation of CHCl2CN
and the other the Payne rearrangement10 to give 10. Both
of these side reactions could perhaps be circumvented by
use of a less basic anion. Thus, we transmetalated alkoxy
anion 8 to its zinc form, creating a weaker base but a
better coordination metal. To our delight, 4a was
obtained as a major product in 65% yield. The imidate
nitrogen anion 11 acts as a good nucleophile to attack
the epoxide, producing 4a . This completes a one-step
epoxide opening in which both a carbon-oxygen and a
carbon-nitrogen bonds are formed.
Guangzhong Wu,* Doris P. Schumacher,
Wanda Tormos, J on E. Clark, and Bruce L. Murphy
Chemical Process Research and Development,
Schering-Plough Research Institute,
2015 Galloping Hill Road, Kenilworth, New J ersey 07033
Received August 1, 1996 (Revised Manuscript Received
February 11, 1997 )
In tr od u ction
Both thiamphenicol (1)1 and its 3′-fluoro derivative,
florfenicol (2)2 are broad spectrum antibiotics possessing
activity against many Gram negative and Gram positive
microorganisms. The current industrial synthesis of
florfenicol is based upon the established thiamphenicol
process which starts with the condensation of p-(meth-
ylsulfonyl)benzaldehyde (14) and glycine.3 The diol
product from the reduction of the corresponding ester is
resolved with unnatural D-(-)-tartaric acid affording
optically pure D-(-)-threo-2-amino-1-(4-(methylsulfonyl)-
phenyl)-1,3-propanediol (3). Various methods were de-
veloped to recycle the L-isomer, but they are generally
lengthy and difficult to operate.4 Since the discovery of
florfenicol in 1979,2 there has been renewed interest in
developing asymmetric syntheses of thiamphenicol based
structures.5 A diastereoselective synthesis of thiam-
phenicol was reported by S. McCombie and T. L.
Nagabhushan.5a However, the low enantioselectivity of
the Sharpless epoxidation on (Z)-allylic alcohol deriva-
tives flaws the approach.6
We now report an improved version of the thiampheni-
col-based synthesis of florfenicol which shortens our
previous route5b by two steps. More importantly, we
report the first enantioselective synthesis of 1′-(R,R)-
(dichloromethyl)oxazoline 5, an intermediate to both
florfenicol and thiamphenicol. In the course of investi-
gating the new synthesis, we have developed a clean
inversion of the benzylic chiral center in 3′-(S,R)-(di-
The synthetic strategy using 4a followed by an inver-
sion was dictated by the fact that the more straightfor-
ward routesdirect epoxidation to a precursor of 4sre-
quired a cis allylic alcohol, and these are not reported to
afford high ee’s.11 To invert 4a , we took advantage of
(1) Cutler, R. A.; Stenger, R. J .; Suter, C. M. J . Am. Chem. Soc. 1952,
74, 5475.
(2) Nagabhushan. T. L. U.S. Patent 4235892, 1980, Schering Corp.;
Chemical Abstracts 1980, 94, 139433.
(3) J acquez, J .; Collet, A.; Wilen, S. In Enatiomers, Racemates, and
Resolutions; J ohn Wiley and Sons: New York, 1981; p 223.
(4) (a) Giordano, C.; Caviccholi, S.; Villa, M. Tetrahedron Lett. 1988,
29, 5561. (b) Giordano, C.; Caviccholi, S.; Levi, S.; Villia, M. J . Org.
Chem. 1991, 56, 6114.
(5) (a) McCombie, S.; Nagabhushan, T. L. Tetrahedron Lett. 1987,
28, 5395. (b) Schumacher, D. P.; Clark, J . E.; Murphy, B. F.; Fischer,
P. A. J . Org. Chem. 1990, 55, 5291. (c) Clark. J . E.; Fischer, P. A.;
Schumacher, D. P. Synthesis 1991, 891. (d) Lou, B.; Zhang, Y.; Dai,
Li. Chem. Ind. 1993, 249.
(7) Takaoka, A.; Iwagiri, H.; Ishikawa, N. Bull. Chem. Soc. J pn.
1979, 52, 3377.
(8) (a) Fotadar, U.; Becu, C.; Borremans, F.; Anteunis, M. J . O.
Tetrahedron 1978, 34, 3537. (b) Sandler, S. R.; Karo, W. Organic
Functional Group Preparations, 2nd ed.; Academic Press: New York,
1989; Chapter 8. (c) Roush, W. R.; Brown, R. J .; Dimane, M. J . Org.
Chem. 1982, 47, 1371.
(9) Schmidt, U.; Respodek, M.; Lieberknecht, A.; Werner, J .; Fischer,
P. Synthesis 1989, 256.
(10) Payne, G. B. J . Org. Chem. 1962, 27, 3819.
(11) Rao, A. V. R.; Rao, P.; Bhanu, M. J . Chem. Soc., Chem.
Commun. 1992, 859.
(6) Katsuki, T.; Sharpless, K. B. J . Am. Chem. Soc. 1980, 102, 5974.
S0022-3263(96)01479-X CCC: $14.00 © 1997 American Chemical Society

Notes
J . Org. Chem., Vol. 62, No. 9, 1997 2997
Sch em e 1
Sch em e 3a
a
(a) Malonic acid, piperidine (0.2 equiv), pyridine, reflux, 2 h.
(b) SOCl2, reflux, 1 h, then NaBH4, EtOH, 5 °C, 1 h. (c) Ti(OPr-i)4
(0.2 equiv), L-DIP (0.2 equiv), t-BuOOH, 4 Å sieves, -20 to -15
°C, 2 h. (d) (i) NaH, THF, 5 °C, 30 min, (ii) ZnCl2 (1.0 equiv), 5 °C,
30 min, (iii) CHCl2CN (1.5 equiv), 55 °C, 16 h. (e) (i) EtSO2Cl,
Et3N, 5 °C, 3 h, (ii) H2SO4, 15 °C, 20 min, (iii) NaOH, 5 °C, 10
min. (f) (i) CF3CHCF2NEt2 (Ishikawa reagent), CH2Cl2, sealed
reflux, 1.5 h, (ii) HOAc, i-PrOH, reflux, 3 h.
Sch em e 2
obtained when 0.2 equiv of the catalyst was used whereas
a 90% ee was achieved with 0.1 equiv of the catalyst.
Following the novel epoxy alcohol opening procedure, 7
was regioselectively opened by a sequential addition of
NaH, ZnCl2, and CHCl2CN to afford oxazoline 4a in 65%
yield. The inversion was carried out according to the
newly-discovered one-pot method to produce 50% of the
desired 1′-(R,R)-oxazoline 5 which is identical to the
authentic sample. The chiral purity was measured to be
greater than 99.9%.
In summary, we have developed a shorter version of
the resolution-based synthesis of florfenicol (2) which has
proven robust enough for commercial use. We have also
developed an enantioselective route to both thiampheni-
col (1) and florfenicol (2).
Exp er im en ta l Section
the neighboring amide group. First, mesylation of the
benzylic alcohol in 4a proceeds nicely to afford 12. A pH
adjustment with dilute H2SO4 hydrolyzes the oxazoline
ring to the open form 13. Re-adjustment of the pH to 10
with aqueous NaOH promotes the intramolecular SN2
attack and produces 1′-oxazoline 5. This transformation
not only inverts the benzylic chiral center but also
isomerizes 4a to the internal oxazoline, exposing the
primary alcohol to the fluorodehydroxylation. A brief
study indicated that EtSO2 gave a 50% overall yield of 5
from 4a while MeSO2 afforded a 45% yield.
With the above developments, we were able to complete
the enantioselective synthesis of 2 as expressed in
Scheme 3. Thus, aldol condensation of 14 with malonic
acid in the presence of pyridine and a catalytic amount
of piperidine followed by dehydration and decarboxylation
gave the trans cinnamic acid derivative 1512 in 83% yield.
A two-step, one-pot procedure was developed for the
reduction of acid 15. First, the acid 15 was converted to
its corresponding acid chloride and then treated with a
solution of NaBH4 in EtOH to produce trans allylic
alcohol 16 in 74% yield. The isolated 16 was subjected
to the Sharpless epoxidation conditions to give the (S,S)-
epoxy alcohol 7 in 82% isolated yield. A 97% ee was
Gen er a l. The 1H and 13C NMR spectra were taken in CDCl3,
and all reactions were carried out under nitrogen unless
otherwise noted. Melting points were not corrected. All starting
materials and reagents were purchased commercially and used
as is without further purifications.
Con d en sa tion of Am in o Diol 3 w ith CHCl2CN to (4R,5R)-
2-(Dich lor om eth yl)-5-[4-(m eth ylsu lfon yl)p h en yl]oxa zole-
4-m eth a n ol (5). To a 500 mL three-neck flask equipped with
a mechanical agitator and a thermometer were sequentially
added at 25 °C 150 mL of 2-propanol, 50.4 g (458 mmol) of
CHCl2CN, 5 mL of concd H2SO4, and 100 g (408 mmol) of amino
diol 3. The mixture was heated to 70 °C for 1.5-2 h to complete
the condensation. The resulting slurry was cooled to 50 °C and
agitated at that temperature for an additional 14 h. The
precipitate was cooled to 5 °C, filtered, and dried at 50 °C to
give 119 g (88%) of 5. The ratio of 5:4 in the isolated solid is
99.5:0.5 as determined by HPLC: [R]23.3D +11.1° (c 5.13, EtOH).
1H NMR (DMSO-d6): δ 8.00 (d, J ) 8.3 Hz, 2H), 7.60 (d, J ) 8.3
Hz, 2H), 7.26 (s, 1H), 5.75 (d, J ) 6.4 Hz, 1H), 5.18 (t, J ) 5.6
Hz, 1H), 4.10-4.05 (m, 1H), 3.75-3.65 (m, 1H), 3.60-3.55 (m,
1H), 3.23 (s, 3H). 13C NMR (DMSO-d6): δ 161.0, 146.0, 140.6,
127.8, 126.1, 83.0, 76.5, 61.8 (2 carbons by Hetcor), 43.5. IR
(KBr, paraffin): 3350, 2930, 1670 cm-1. Mp: 144-145 °C. Anal.
Calcd for C12H13Cl2NO4S: C, 42.63; H, 3.88; N, 4.14. Found:
C, 42.29; H, 4.11; N, 4.24. Analyses for the minor product 4:
[R]22.8 -150.3° (c 4.79, DMSO). 1H NMR (DMSO-d6): δ 7.81
(d, J ) 8.2 Hz, 1H), 7.56 (d, J ) 8.2 Hz, 1H), 6.95 (s, 1H), 5.83
(d, J ) 4.8 Hz, 1H), 4.81 (t, J ) 4.0 Hz, 1H), 4.56-4.51 (m, 1H),
4.44-4.33 (m, 2H), 3.13 (s, 3H, Me). 13C NMR (DMSO-d6): δ
D
(12) Organic Reactions, Vol. 15, p 268.

2998 J . Org. Chem., Vol. 62, No. 9, 1997
Notes
162.3, 148.4, 139.7, 127.8, 126.7, 71.8, 71.5, 70.7, 62.3, 43.9. IR
on a Chiralcel OJ column with hexane:2-propanol:acetonitrile
(50:50:1) as the mobile phase. The (R,R)-isomer was also
prepared by following the same procedure. 1H NMR: δ 7.92 (d,
J ) 8 Hz, 2H), 7.49 (d, J ) 8 Hz, 2H), 4.10 (dd, J ) 17, 3 Hz,
1H), 4.04 (s, 1H), 3.85 (dd, J ) 17, 3 Hz, 1H), 3.20-3.17 (m,
1H), 3.05 (s, 3H), 1.88 (bs, 1H). 13C NMR: δ 143.4, 140.2, 127.6,
126.5, 62.9, 60.7, 54.5, 44.5. Mp: 104-106 °C. IR (paraffin):
3300, 2940, 1600 cm-1. Anal. Calcd for C10H12O4S: C, 52.62;
H, 5.30; S, 14.04. Found: C, 52.60; H, 5.27; S, 14.36.
(KBr, paraffin): 3250, 2950, 1660 cm-1
. Mp: 12.5-126.5 °C.
Anal. Calcd for C12H13Cl2NO4S: C, 42.63; H, 3.88; N, 4.14.
Found: C, 42.48; H, 4.04; N, 4.22.
Con ver sion of 5 to F lor fen icol (2). To a 30 mL nonstirred
Teflon tube were added 8.5 g of a 23% CH2Cl2 solution of CF3-
CHCF2NEt2 (Ishikawa reagent, 9 mmol) and 2.0 g (6 mmol) of
5. The Teflon tube was inserted into an autoclave. The
autoclave was sealed, heated to 100 °C in an oil-bath for 1-3 h,
and cooled with an ice bath. The content was transferred into
a 250 mL flask for hydrolysis. To the flask were added 0.35 g
of KOAc to pH 5 and 2 mL of MeOH. Most of solvent in the
flask was removed under vacuum. To the residue was added a
20 mL mixture of 2-propanol (IPA) and water (65:35). The IPA-
water mixture was refluxed for 3 h to complete the hydrolysis.
After removal of about half of the solvent under vacuum, the
product began to crystallize. The precipitate was cooled to 5
°C, filtered, and dried at 50 °C to afford 2.0 g of florfenicol (2).
Recrystallization from 2-propanol and water gave the final drug
in >98% purity. The assays of the product obtained from this
route are identical to that reported.5b
(3S,4R)-2-(Dich lor om eth yl)-4,5-d ih yd r o-R-[4-(m eth ylsu l-
fon yl)p h en yl]oxa zole-4-m eth a n ol (4a ). To a 1 L three-neck
oven-dried flask with a condenser and a mechanical agitator
were added sequentially 6.1 g (153 mmol) of NaH (60%) and 90
mL of THF. To the cooled suspension at 5 °C was added
dropwise 300 mL of a THF solution of 30 g (128 mmol) of (S,S)-
3-[4-(methylsulfonyl)phenyl]-2,3-epoxy alcohol 7. The resulting
mixture was agitated at 5 °C for 30 min to form the sodium
anion. To the above solution was added 17.8 g (128 mmol) of
dry ZnCl2 dissolved in 250 mL of THF. After 30 min of agitation
at 5 °C, 17.0 g (153 mmol) of CHCl2CN in 10 mL of THF and 1
g of 4 Å sieves was added dropwise into the flask. The contents
were heated to 55 °C for 16 h, cooled to 25 °C, and quenched
with aqueous NaHCO3. The product was extracted with 4 ×
400 mL of EtOAc, washed with brine, and concentrated under
vacuum. The product solidified after addition of 50 mL of IPA
and was filtered and dried at 55 °C to give 16.0 g of 4a . The
solution yield was 65% as assayed by HPLC vs an external
standard. The crude product was recrystallized from methyl
(E)-3-[4-(Meth ylsu lfon yl)p h en yl]p r op en ic Acid (15). To
a three-neck 3 L round bottom flask equipped with a condenser,
a thermometer, and a mechanical agitator were added sequen-
tially 596 g (2.99 mol) of malonic acid, 506 mL of pyridine, 30
mL of piperidine, and 300 g (1.49 mol) of p-(methylsulfonyl)-
benzaldehyde 14. The mixture was heated to 95-100 °C for 4
h, cooled to 25 °C, and quenched slowly into 3 L of an ice-HCl
solution. The precipitate was filtered and dried at 50 °C to give
340 g of 15 as white crystals. The yield was 83% after correction
for the purity. 1H NMR (DMSO-d6): δ 7.90 (d, J ) 8.9 Hz, 2H),
7.87 (d, J ) 8.9 Hz, 2H), 7.62 (d, J ) 16 Hz, 1H), 6.65 (d, J ) 16
Hz, 1H), 3.19 (s, 3H, Me). 13C NMR (DMSO-d6): δ 167.5, 142.2,
141.8, 139.5, 129.2, 127.8, 123.1, 43.7. IR (KBr, paraffin): 2940,
1680 cm-1. Mp: 294-296 °C. Anal. Calcd for C10H10H4S: C,
53.08; H, 4.45; S, 14.17. Found: C, 52.78; H, 4.66; S, 13.94.
(E)-[4-(Meth ylsu lfon yl)p h en yl]p r op en ol (16). To an oven-
dried 250 mL three-neck flask was charged 96 mL (1.35 mol) of
SOCl2 and 50 g (0.225 mol) of (E)-3-[4-(methylsulfonyl)phenyl]-
propenic acid (15). The reaction mixture was heated to reflux
for 1 h. The excess SOCl2 was removed via azeotropic distillation
with CH2Cl2, and the residue was made into a 100 mL CH2Cl2
solution. This solution was then added dropwise to precooled
(5 °C) 42 g (1.1 mol) of NaBH4 in EtOH. The resulting mixture
was agitated at 10 °C for 1 h, quenched slowly into an ice-HCl
solution, and extracted with 3 × 300 mL of CH2Cl2. The product
began to crystallize after concentration of the organic layer. The
precipitate was filtered and dried at 50 °C to give 28 g of 16
(74% yield) with a solution yield of 94%. 1H NMR: δ 7.80 (d, J
) 8.2 Hz, 2H), 7.62 (d, J ) 8.2 Hz, 2H), 6.62 (d, J ) 16.2 Hz,
1H), 6.55 (dt, J ) 16.2, 3.8 Hz, 2H), 4.96 (t, J ) 5.0 Hz, 1H,
OH), 4.12-4.10 (m, 1H), 3.14 (s, 3H, Me). 13C NMR (DMSO-d6)
δ 142.3, 139.2, 135.5, 127.7, 127.0, 126.8, 61.5, 43.9. IR
isobutyl ketone to afford 4a with a 97% purity. [R]22.8 -43.8°
D
(c 4.69, DMSO). 1H NMR: δ 7.93 (d, J ) 8.4 Hz, 2H), 7.60 (d,
J ) 8.4 Hz, 2H), 6.26 (s, 1H), 5.14 (d, J ) 4.0 Hz, 1H), 4.57 (ddd,
J ) 8.7, 7.9, 4.0 Hz, 1H), 4.46 (dd, J ) 9.7, 7.9 Hz, 1H), 4.28
(dd, J ) 9.7, 8.7 Hz, 1H), 3.05 (s, 3H), 2.60 (bs, 1H). 13C (DMSO-
d6): δ 162.4, 148.9, 139.9, 127.7, 127.1, 72.6, 72.1, 70.2, 62.4,
43.8. IR (KBr, (paraffin): 3240 (OH), 2900, 1650 cm-1
. Mp:
156.5-157.5 °C. Anal. Calcd for C12H13Cl2NO4S: C, 42.63; H,
3.88; N, 4.14. Found: C, 42.39; H, 4.16; N, 4.15.
(4R ,5R )-2-(D ic h lo r o m e t h y l)-5-[4-(m e t h y ls u lfo n y l)-
p h en yl]oxa zole-4-m eth a n ol (5). To an oven-dried 100 mL
flask were charged at 25 °C sequentially 3.5 g (10 mmol) of 4a ,
5 mL of pyridine, and 2.8 mL of Et
3N. To the cooled mixture at
5 °C was added dropwise 0.95 mL (12 mmol) of MeSO2Cl. The
resulting solution was stirred at 5 °C for 2 h, and the pH was
then adjusted to 2.0 with 3.0 N H2SO4. Then 5 mL of THF was
added to make the hydrolysis homogeneous. The reaction was
warmed to 25 °C for 10 min and treated with 50% NaOH to pH
12.5 to complete the cyclization. The reaction mixture was
extracted with 3 × 40 mL of EtOAc, washed with brine, and
concentrated to give crude 5. An analytical sample was purified
by flash column chromatography. The solution yield was 40-
45%, and a 50% yield was obtained when EtSO2Cl was used.
The ee was determined to be >99.9% on a J . T. Baker Chiralcel
OJ column with hexane:2-propanol:acetonitrile (69:30:1) as the
mobile phase. The assays of this product are identical to that
of the authentic sample.
((paraffin): 3520, 3340, 2930, 1420 cm-1
Anal. Calcd for C10H12O3S: C, 56.59; 5.70; S, 15.1. Found: C,
56.36; H, 5.70; S, 15.1.
. Mp: 126-127 °C.
(S,S)-3-[4-(Meth ylsu lfon yl)p h en yl]-2,3-ep oxyp r op yl Al-
coh ol (7). To an oven-dried 500 mL three-neck flask with a
thermometer and a mechanical agitator at -20 °C were added
sequentially 8 g of 4 Å molecular sieves powder, 1.74 g (7.4
mmol) of diisopropyl L-tartrate, and 2.16 g (7.4 mmol) of Ti(OPr-
i)4. The reaction mixture was agitated at -20 °C for 30 min. To
the resulting mixture were added dropwise 8.1 g (37 mmol) of 8
dissolved in 500 mL of CH2Cl2 and 13.4 mL of 3.0 M t-BuOOH
in 2,2,4-trimethylpentane. The reaction was stirred at -20 °C
for 4 h, quenched with 6.0 mL of Me2S, and filtered. After
addition of 250 mL of saturated NaF, the filtrate was stirred at
ambient temperature for 16 h, filtered through a pad of Celite,
extracted with 3 × 100 mL of CH2Cl2, and washed with 2 × 100
mL of water. After concentration of the organic layer, the
precipitate was filtered and dried at 50 °C to give 7.15 g (82%
yield) of 7 as white crystals. The ee was determined to be 99.9%
Ack n ow led gm en t. We thank the staff of the Physi-
cal-Analytical Department for high-resolution mass
spectral (HRMS) data, microanalyses, and rotations. We
also acknowledge the personnel of the pilot plant for the
successful scale up.
Su p p or tin g In for m a tion Ava ila ble: 1H NMR, 13C NMR,
and IR spectra for 2, 4, 4a , 5, 7, 15, and 16, also chiral HPLC
chromatograms for 5 and 7 (26 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
J O961479+