A scalable synthesis of 2S-hydroxymutilin via a modified rubottom oxidation

Article abstract of DOI:10.1021/jo801969e

(Chemical Equation Presented) A scalable synthesis of 2S-hydroxymutilin from pleuromutilin was developed. The synthesis is highlighted by the Rubottom oxidation of a silyl enol ether, conducted in the presence of acetic acid and pyridine, which allows for

Full text of DOI:10.1021/jo801969e

A Scalable Synthesis of 2S-Hydroxymutilin via a
Modified Rubottom Oxidation
Huan Wang,* Yemane W. Andemichael, and
Frederick G. Vogt
Chemical DeVelopment, GlaxoSmithKline, 709 Swedeland
Road, King of Prussia, PennsylVania 19406
FIGURE 1. Pleuromutilin, 2S-hydroxymutilin, and mutilin.
SCHEME 1. Proposed Route to 2S-Hydroxymutilin
huan.2.wang@gsk.com
ReceiVed September 9, 2008
A scalable synthesis of 2S-hydroxymutilin from pleuromutilin
was developed. The synthesis is highlighted by the Rubottom
oxidation of a silyl enol ether, conducted in the presence of
acetic acid and pyridine, which allows for an efficient and
selective oxidation.
product was obtained, the yield was unacceptably low (<30%);
therefore, this three-step sequence was examined in greater
detail.
A closer look at the silyl enol ether formation showed that
under the initial reaction conditions (TMSOTf/Et3N, Scheme
2), the desired 2,3-enol ether 4 was not formed exclusively.
Regioisomeric 3,4-enol ether 7 was also formed (4/7 ca. 15:1).
The product mixture also appeared unstable: when stored as an
oil in a refrigerator overnight, the ratio of the two enol ethers
deteriorated to ∼2.3:1 with a small amount of hydrolyzed ketone
8. The ratio of the three was also very inconsistent and
dependent on the workup solvent and procedure.
Pleuromutilin (1, Figure 1) is a naturally occurring antibiotic
isolated from Pleurotus mutilus and other basidiomycetes.1 The
derivatives of pleuromutilin have received considerable attention
in recent years because they fulfill the need for new antibiotics
with a distinctly novel mode of action.2 Derivatives of 2S-
hydroxymutilin (2, Figure 1) have been identified as potent
antimicrobial agents3 and our interest in this class of compounds
has necessitated a method to prepare 2S-hydroxymutilin that is
amenable to large-scale synthesis.4
A straightforward approach to 2S-hydroxymutilin at the outset
seemed to be R-hydroxylation of the cyclopentanone moiety of
mutilin 3 via an oxidation of its silyl enol ether 4,5 with the
oxidation expected to occur from the less hindered R face
(Scheme 1).6 However, during our early efforts, this transforma-
tion often resulted in complex mixtures. When the desired
1H revealed the presence of Et3N+H in crude 4, leading to
the hypothesis that this migration was catalyzed by a trace
amount of acid in the product.7 Thus, the enol ether migration
was circumvented by using a nonacidic procedure: deprotonation
of 3 with LiHMDS provided exclusively the kinetic enolate,
and silylation with TMSCl provided the 2,3-silyl enol ether 4
in nearly quantitative yield after aqueous workup. The product
was stable for months when stored cold, or even at room
temperature as an oil.
Observations in the Rubottom oxidation8 of 4 further
demonstrated the lability of the C4 proton in the presence of a
neighboring carbocation. When silyl enol ether 4 was treated
with mCPBA9 in hexane, only a small amount of the desired
* To whom correspondence should be addressed. Phone: 610-270-5362. Fax:
610-270-4022.
(1) (a) Kavanagh, F.; Hervey, A.; Robbins, W. J. Proc. Natl. Acad. Sci. U.S.A.
1951, 37, 570. (b) Knauseder, F.; Brandl, E. J. Antibiot. 1976, 29, 125.
(2) Hunt, E. Drugs Future 2000, 25, 1163, and references cited therein.
(3) Brooks, G.; Hunt, E. PCT Int. Appl. WO 2001074788 A1, 2001.
(4) (a) 2S-Hydroxymutilin has been prepared via O-H insertion of 2-dia-
zomutilin: see ref 3. However, large-scale application of this method is limited
by the high cost of the diazotization reagent and the stability of the intermediates.
(b) 2S-Hydroxymutilin has been obtained in the microbial hydroxylation of
mutilin, see: Hanson, R. L.; Matson, J. A.; Brzozowski, D. B.; LaPorte, T. L.;
Springer, D. M.; Patel, R. N. Org. Process Res. DeV. 2002, 6, 482.
(5) Direct oxidation of the enolate was also investigated. The oxidation with
molecular oxygen often suffered from overoxidation to the diketone. Oxidation
of the enolate with Davis oxaziridine was successful on a small scale, but was
not deemed as a long-term option due to availability and safety concerns of the
reagent.
(6) For the X-ray structure of mutilin see: (a) Dobler, M.; Du¨rr, B. G. Cryst.
Struct. Commun. 1975, 4, 259. (b) Pilati, T. Acta Crystallogr. 1995, C51, 2676.
For a brief discussion of its conformation, see: (c) Boeckman, R. K.; Springer,
D. M.; Alessi, T. R. J. Am. Chem. Soc. 1989, 111, 8284.
(7) Emde, H.; Domsch, D.; Feger, H.; Frick, U.; Go¨tz, A.; Hergott, H. H.;
Hofmann, K.; Kober, W.; Kra¨geloh, K.; Oesterle, T.; Steppan, W.; West, W.;
Simchen, G. Synthesis 1982, 1.
(8) For a general review on the Rubottom oxidation, see: Wolfe, J. P.
Rubottom oxidation In Name Reactions for Functional Group Transformations;
Li, J. J., Corey, E. J., Eds.; Wiley: New York, 2007, pp 282-290.
478 J. Org. Chem. 2009, 74, 478–481
10.1021/jo801969e CCC: $40.75 2009 American Chemical Society
Published on Web 11/21/2008

SCHEME 2. Migration of the 2,3-Enol Ether
SCHEME 4. Formation of Overoxidation Products
SCHEME 3. Oxidation of 4 under Acidic Conditions
SCHEME 5. Oxidation in CH2Cl2 with NaHCO3 Buffer
product 12 was observed. The major product obtained appeared
to be oxidized enol ether 11, presumably through an acid-
catalyzed epoxide opening followed by loss of the C4 proton
(Scheme 3).
This problem was initially overcome by buffering the reaction
system with NaHCO3. The oxidation performed in the buffered
systems provided the desired product, although the reaction was
slow in several of the solvents screened (hexane, TBME, and
toluene). The main side products appeared to be ketone 8 from
hydrolysis of the silyl enol ether, as well as small amounts of
overoxidation products. The latter was particularly problematic
as attempts to drive the slow reaction to completion by additional
mCPBA often led to more overoxidation products. These side
products and the mechanism of their formation are shown in
Scheme 4: the persistence of the intermediary siloxycarbocation
10 not only led to the normal double oxidation products (15
and 18),10 the oxonium ion at C3 in intermediate 17 also
underwent an intramolecular hydride transfer from C11 to give
2,3,4-triol 19.11
as enones 20 and 21 (Scheme 5). The exact mechanism for the
formation of the two enones is unclear at the moment; however,
the oxidation of C11 in 21 presumably proceeded by the
intramolecular hydride transfer mechanism similar to that in the
formation of 19, which reinforced the notion that the persistence
of the C3 oxonium ion intermediate 10 was detrimental to
the efficiency of the reaction.
With a NaHCO3 buffer, THF proved to be the best solvent,
and provided complete reaction overnight. The epoxidation of
the C12 olefin was also completely suppressed,12 affording the
desired product in good yield. However, difficulty was encoun-
tered when scale-up was attempted. The heterogeneous reaction
suffered from poor reproducibility, giving significantly higher
amounts of hydrolysis product 8 when performed at multigram
scale. There was also growing concern about the compatibility
of THF and mCPBA from a safety standpoint.13 Therefore,
another solvent system was needed for the oxidation reaction.
The reaction was vastly improved by using a mixture of acetic
acid and CH2Cl2 as the reaction solvent. Consistent with
previous observations, the oxidation in CH2Cl2 was rapid even
at -20 °C. However, the addition of acetic acid led to a much
cleaner reaction profile, affording a high-yielding R-hydroxy-
lation of the cyclopentanone moiety (entry 3, Table 1). Addition
of pyridine in this solvent mixture further improved the profile
and led to a slightly higher yield (enrty 4).
Among all the solvents screened in the NaHCO3-buffered
oxidation, CH2Cl2 provided by far the fastest reaction. However,
a complex reaction profile seriously detracted from the increased
reaction rate. Several side products were observed, including
(after desilylation) the hydrolysis product 3, epoxides resulting
from epoxidation of the terminal olefin moiety on C12, as well
(9) The Rubottom oxidation was also attempted with in situ generated
peroxyimidic acid and dioxirane. Unfortunately, hydrolysis of the silyl enol ether
became a serious side reaction in both cases. A significant amount of olefin
epoxidation was also observed when methyl(trifluoromethyl)dioxirane was used
as the oxidant.
(10) For an example of double oxidation under Rubottom conditions, see:
Horiguchi, Y.; Nakamura, E.; Kuwajima, I. Tetrahedron Lett. 1989, 30, 3323.
(11) The acid-catalyzed rearrangement from mutilin to the 4-epimutilin
skeleton is well known: Berner, H.; Schulz, G.; Schneider, H. Tetrahedron 1980,
36, 1807.
(12) This was confirmed by the observation that the C12 olefin in 8 did not
undergo epoxidation when exposed to NaHCO3-buffered mCPBA in THF.
(13) Bretherick’s Handbook of ReactiVe Chemical Hazards; Urben, P. G.,
Ed.; Butterworth-Heinemann: Burlington, MA, 1999.
J. Org. Chem. Vol. 74, No. 1, 2009 479

TABLE 1. Buffer Effect on the Rubottom Oxidation
SCHEME 7. Scalable Synthesis of 2S-Hydroxymutilin
buffer/solvent
conditions
hydrolysis yields,a
%
1
2
3
4
NaHCO3/THF
NaHCO3/CH2Cl2 -10 to 0 °C, 2 h
HOAc/CH2Cl2
-10 °C to rt, overnight trace to 34% 42-75
b
b
-
-
-10 °C, 1.5 h
trace
trace
69
78
Py/HOAc/CH2Cl2 -10 °C, 1.5 h
a Isolated yields. b Complex mixture, see text.
After the oxidation, the desilylation of the O11 and O14 TMS
groups with use of TBAF provided the desired product, though
it was slow and often led to an unidentified complex mixture.
Desilylation under acidic conditions provided a faster and much
cleaner transformation: when treated with HCl, the O11 TMS
was desilylated within 5-10 min, while the more hindered O14
TMS was removed in 1-2 h.
SCHEME 6. Plausible Mechanism for Additive Effect
This synthetic sequence was performed successfully at a
kilogram scale. Mutilin 3 was obtained from hydrolysis of
pleuromutilin18 and was converted into silyl enol ether 4
smoothly by using LHMDS/TMSCl. The Rubottom oxidation
was carried out by controlled addition of 4 into a cold mixture
of acetic acid, pyridine, and mCPBA. It proved very important
to have efficient cooling to maintain the low reaction temperature
necessary for the high selectivity, as well as to minimize
consumption of mCPBA by pyridine during the time of addition.
After desilylation and crystallization, 2S-hydroxymutilin 2 was
obtained in 78% yield (Scheme 7).19
In summary, the Rubottom oxidation of silyl enol ether 4
was studied in detail. The lability of H4 in the presence of a
neighboring carbocation proved detrimental to the efficiency
of the reaction. This reaction was improved by buffering the
reaction with an inorganic base, or, more interestingly, acetic
acid. The role of acetic acid is presumably to minimize the
concentration of oxonium intermediate 10. This modified
procedure with acetic acid allowed for an efficient and selective
oxidation of 4 and a synthesis of 2S-hydroxymutilin from
pleuromutilin on a kilogram scale.
While NaHCO3 and other inorganic bases are routinely used
to buffer the Rubottom reaction involving acid-sensitive sub-
strates, acetic acid has not been widely used as an additive in this
reaction. However, both seem to facilitate the Rubottom reaction
of silyl enol ether 4. A plausible explanation of their different roles
in this reaction is shown in Scheme 6. A basic buffer would inhibit
the formation of 10 from the presumed silyloxyoxirane intermediate
9,14 while favoring the epoxide opening initiated by a direct
nucleophilic attack on the TMS group (9 to 12). Addition of HOAc,
on the other hand, would accelerate the formation of 10, but it
would also trap the latter to form 22,15 which then collapses to
yield the desired product 12. In this case, both a basic and an acidic
additive help in suppressing the buildup of oxonium intermediate
10,16 which is responsible for several of the side reaction pathways
described above.17 The large amount of acetic acid in the reaction
system can also minimize the concentration of 10 by cleaving the
O-Si bond, although the TMS cleavage of silyl enol ether 4 is
clearly not as fast as its oxidation as only trace amounts of
hydrolysis were observed. This is particularly remarkable consider-
ing the high acid-sensitivity and migratory tendency of silyl enol
ether 4.
Experimental Section
General. Unless otherwise indicated, all reactions were con-
ducted in glass-lined reactors under a nitrogen atmosphere. Solvents
and reagents were obtained from commercial sources and used
without further purification. Pleuromutilin 1 was obtained via
fermentation.
Mutilin 3. A mixture of pleuromutilin 1 (1.2 kg, 3.17 mol) and
NaOH (50% w/w, 963 g, 12 mol, 3.8 equiv) in ethanol (7.2 L) and
water (4.5 L) was stirred at 50 °C for 3 h. The mixture was
concentrated. The crystalline product was isolated by filtration,
washed with water (1.7 L × 2) and heptane (0.7 L), and then dried
under vacuum to provide 740 g (2.31 mol, 73%) of product. ES-
MS: m/z 285 (M - 2H2O + H+), 303 (M - H2O + H+), 343 (M
+ Na+), 384 (M + Na+ + CH3CN). 1H NMR (CDCl3, 300.13 MHz)
δ (ppm) 6.11 (dd, J ) 17.8, 11.1 Hz, 1H), 5.28-5.41 (m, 2H),
4.36 (dd, J ) 7.3, 5.8 Hz, 1H), 3.42 (t, J ) 6.6 Hz, 1H), 2.13-2.32
(14) (a) Dodd, J. H.; Starrett, J. E.; Weinreb, S. M. J. Am. Chem. Soc. 1984,
106, 1811. (b) Paquette, L. A.; Lin, H.-S.; Gallucci, J. C. Tetrahedron Lett. 1987,
28, 1363. (c) Gleiter, R.; Staib, M.; Ackermann, U. Liebigs Ann. 1995, 1655.
(15) (a) The trapping of silyloxy carbocation with MCBA has been
reported: Hassner, A.; Reuss, R. H.; Pinnick, H. W. J. Org. Chem. 1975, 40,
3427. (b) For a related example see: Rubottom, G. M.; Gruber, J. M.; Boeckman,
R. K.; Ramaiah, M.; Medwid, J. B. Tetrahedron Lett. 1978, 19, 4603.
(16) For evidence in the intermediacy of the siloxycarbocation, see ref
14b.
(17) An intramolecular pathway is also possible: as the bulky silyl group
presumably resides on the less crowded R face, the acetate could attack the
oxonium ion 10 from the ꢀ-face and the resultant stereoisomer would undergo
a 1,4-silicon shift to generate the R-hydroxyketone moiety. However, the
corresponding 2-silyloxy product was never isolated.
(18) Pleuromutilin 1 was obtained from fermentation.
(19) The process could be further streamlined by using propyl acetate as a
single solvent through oxidation, desilylation, and isolation, allowing the entire
sequence to be conducted in one reaction vessel without formal isolation of the
intermediates, while providing 2S-hydroxymutilin in 58% overall yield.
480 J. Org. Chem. Vol. 74, No. 1, 2009

1
(m, 3H), 2.05 (m, 1H), 1.88-1.96 (dd, J ) 15.9, 7.8 Hz, 1H),
1.42-1.78 (m, 7H), 1.37 (s, 3H), 1.16 (s, 3H), 1.08-1.18 (m, 1H),
0.98 (d, J ) 7.0 Hz, 3H), 0.94 (d, J ) 7.2 Hz, 3H). 13C NMR
(CDCl3, 100.61 MHz) δ (ppm) 217.7, 139.4, 115.8, 75.0, 66.7, 59.0,
45.3, 45.2, 45.0, 42.3, 36.8, 36.4, 34.4, 30.3, 28.6, 27.1, 25.0, 18.1,
13.4, 11.2.
+ Na+), 544 (M + Na+ + CH3CN). H (CDCl3, 400.13 MHz) δ
(ppm) 6.16 (dd, J ) 17.6, 11.2 Hz, 1H), 5.12-5.40 (m, 2H), 4.43
(d, J ) 8.1 Hz, 1H), 4.00 (td, J ) 8.6, 2.3 Hz, 1H), 3.44 (d, J )
6.0 Hz, 1H), 3.02 (br, 1H), 2.25-2.34 (m, 1H), 2.21 (m, 1H), 2.07
(dd, J ) 13.0, 8.7 Hz, 1H), 1.87-1.99 (m, 2H), 1.23-1.56 (m,
5H), 1.35 (s, 3H), 1.08 (s, 3H), 0.94-1.12 (m, 1H), 0.88 (m, 6H),
0.13 (s, 9H), 0.12 (s, 9H). 13C NMR (CDCl3, 100.61 MHz) δ (ppm)
219.5, 140.8, 116.2, 77.1, 72.5, 67.4, 57.4, 47.4, 45.0, 43.8, 42.1,
38.8, 36.3, 35.1, 33.0, 29.0, 26.9, 18.2, 14.8, 12.4, 1.3, 1.0.
The resulting solution of hydroxyketone 12 was cooled to 5-10
°C and treated with 1.4 N HCl/MeOH (2.4 L, prepared from conc
HCl and MeOH). The reaction mixture was stirred until complete
hydrolysis was achieved (∼1 h). The reaction mixture was quenched
with saturated aq NaHCO3 (1.8 L) and concentrated to remove ∼7
L of solvents, during which the product started to precipitate. To
the vigorously stirred mixture was added water (1.8 L). The
resulting slurry was stirred for 0.5 h and filtered. The residue was
washed with water (1.35 L) and heptane (0.9 L). The wet cake
was dried in a vacuum oven at 45 °C overnight. The dry white
solid product weighed 578 g and contained ∼8% of mutilin 3.
Further purification and crystallization of 2-hydroxymutilin 2 was
demonstrated on a 50 g scale: to a stirred mixture of crude product
(50 g) in 50 mL of THF at 55-60 °C was slowly added 750 mL
of cyclohexane. After being stirred for ∼1 h at 60 °C, the mixture
was allowed to cool slowly to rt. The mixture was stirred for 1 h
and filtered, then the residue was washed with cyclohexane (250
mL). The product was dried under vacuum: 38 g, 78% (from mutilin
3) and free of 3. ES-MS: m/z 301 (M - 2H2O + H+), 359 (M +
Na+), 400 (M + Na+ + CH3CN). 1H (CDCl3, 400.13 MHz) δ (ppm)
6.12 (dd, J ) 17.9, 11.2 Hz, 1H), 5.27-5.38 (m, 2H), 4.34 (d, J )
7.7 Hz, 1H), 3.99 (t, J ) 8.6 Hz, 1H), 3.39 (d, J ) 6.3 Hz, 1H),
2.20 (m, 1H), 2.05-2.15 (m, 2H), 1.89-1.97 (m, 2H), 1.66 (d, J
) 15.9 Hz, 1H), 1.34-1.54 (m, 4H), 1.39 (s, 3H), 1.15 (s, 3H),
0.95-0.99 (m, 7H). 13C NMR (CDCl3, 100.61 MHz) δ (ppm) 218.8,
139.5, 115.7, 74.6, 72.2, 66.8, 57.4, 45.34, 45.30, 42.8, 42.0, 38.1,
36.6, 34.8, 32.5, 28.7, 26.9, 18.1, 13.8, 11.6.20
Silyl Enol Ether 4. A solution of mutilin 3 (300 g, 0.936 mol,
1 equiv) and trimethylsilyl chloride (426 mL, 3.37 mol, 3.6 equiv)
in THF (3 L) was cooled to -18 °C. LiHMDS in hexanes (1.0 M,
3.1 L, 3.3 equiv) was slowly added while maintaining the internal
temperature below -8 °C. After 2 h of stirring the reaction was
deemed complete by TLC. The crude mixture was diluted with
cyclohexane (1.5 L) and then quenched with saturated aq NaHCO3
(1.5 L). The organic layer was washed with saturated aq NaHCO3
(1.0 L), concentrated, and used in the next step without further
purification (quantitative yield assumed). An analytically pure
sample can be obtained after flash column chromatography: 1H (d8-
toluene, 400.13 MHz) δ (ppm) 6.35 (dd, J ) 17.37, 11.16 Hz, 1
H), 5.35 (dd, J ) 17.57, 1.86 Hz, 1 H), 5.25 (dd, J ) 11.17, 1.65
Hz, 1 H), 4.69 (d, J ) 7.86 Hz, 1 H), 4.50 (s, 1 H), 3.86 (d, J )
6.62 Hz, 1 H), 2.83 (s, 1 H), 2.28-2.46 (m, 1 H), 2.10-2.20 (m,
1 H), 1.81-1.92 (m, 1 H), 1.72-1.80 (m, 1 H), 1.63-1.72 (m, 1
H), 1.53-1.62 (m, 2 H), 1.51 (s, 2 H), 1.40 (s, 2 H), 1.31-1.39
(m, 1 H), 1.09 (d, J ) 7.03 Hz, 2 H), 0.94 (d, J ) 7.03 Hz, 2 H),
0.21 (s, 9 H), 0.19 (s, 9 H), 0.12 (s, 9 H). 13C NMR (d8-toluene,
100.62 MHz) δ (ppm) 159.0, 142.6, 116.2, 100.3, 78.5, 69.0, 53.1,
48.6, 48.5, 45.6, 43.5, 39.4, 38.0, 36.7, 33.0, 29.8, 28.6, 19.4, 17.2,
14.1, 2.0, 1.4, 0.4.
2S-Hydroxymutilin 2. A mixture of mCPBA (∼77%, 562 g,
2.51 mol, 1.5 equiv) and DCM (9.0 L) was cooled to -25 °C with
stirring. To the mixture was added acetic acid (2.25 L, 39.3 mol,
23 equiv) followed by pyridine (0.743 L, 9.19 mol, 5.5 equiv) while
maintaining the internal temperature at less than -20 °C. To the
cold mixture was slowly added a solution of silyl enol ether 4 (900
g, 1.68 mol) in DCM (2.7 L) over 100 min while maintaining the
internal temperature at less than -20 °C. After ∼0.5 h of stirring,
the reaction temperature was raised to -10 °C over 75 min and
the reaction was quenched with 10% aq Na2SO3 (1.8 L). After 0.5 h
of stirring at rt, the organic layer was separated and then washed
twice with water (2 × 3.6 L). The organic layer was concentrated
to ∼4 L. To the residue was added ethyl acetate (4.5 L) and the
mixture was concentrated to ∼4 L. The crude mixture was diluted
with cyclohexane (4.5 L) and then washed with saturated aq
NaHCO3 (2 × 3.6 L). The crude hydroxyketone 12 was used
directly in the next step, but an analytically pure sample can be
obtained after flash column chromatography: ES-MS: m/z 503 (M
Supporting Information Available: Experimental details,
characterizations, and 1H, 13C, and 2D NMR spectra for all new
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
JO801969E
(20) For
a detailed spectroscopic and computational study of 2S-hy-
droxymutilin, see: Vogt, F. G.; Spoors, G. P.; Su, Q.; Andemichael, Y. W.; Wang,
H.; Potter, T. C.; Minick, D. J. J. Mol. Struct. 2006, 797, 5.
J. Org. Chem. Vol. 74, No. 1, 2009 481