A streamlined synthesis of androstadiene C-17 ester derivatives

Article abstract of DOI:10.2533/chimia.2011.877

The development of a fully telescoped synthesis of a derivative of androstadiene C-17 esters made from epoxyparamethasone was demonstrated. This streamlining allowed for the elimination of isolation and solvent change after each synthetic step. Thus it not only drastically reduced the solvent waste, but also minimized the potential exposure to highly active intermediates thereby increasing the overall yield. The intuitively obvious advantage inherent to lowering the number of solvents was illustrated by applying standard green metrics.

Full text of DOI:10.2533/chimia.2011.877

Columns
CHIMIA 2011, 65, No. 11 877
doi:10.2533/chimia.2011.877
International Year of Chemistry 2011
Popular Scientific Papers
Presented by the Division of Industrial Chemistry
2-carboxylate moiety was used to cap the free hydroxyl group to
the desired product 1, ready for final processing to the targeted
solid state form.
A Streamlined Synthesis of Androstadiene C-17 Ester
Derivatives
The highly potent nature of this class of compounds prompted us
to consider a fully streamlined process with minimal number of
operations and open handling of solids. This is of particular ad-
vantage for large-scale processes when considered from health,
safety and environmental perspectives. The present article de-
scribes our progress in this direction.
Fabrice Gallou*, Manuela Seeger-Weibel, and
Daniel Lupp
*Correspondence: Dr F. Gallou, Chemical and Analytical Development, Novartis
Pharma AG, CHBS, WSJ-145.7.52, Novartis Campus, CH-4056 Basel
Tel.: +41 61 324 99 11, E-mail: fabrice.gallou@novartis.com
Abstract: The development of a fully telescoped synthesis of a
derivative of androstadiene C-17 esters made from epoxypara-
methasone was demonstrated. This streamlining allowed for the
elimination of isolation and solvent change after each synthetic
step. Thus it not only drastically reduced the solvent waste, but
also minimized the potential exposure to highly active intermedi-
ates thereby increasing the overall yield. The intuitively obvious
advantage inherent to lowering the number of solvents was il-
lustrated by applying standard green metrics.
Results
Oxidative Cleavage
Our work started with an extensive screening of conditions for the
selective oxidative cleavage of the α-hydroxy ketone 2 (Scheme
2). Periodic acid was shown to be optimal for this transforma-
tion. Other catalytic methods led to, at best, mediocre selectivity
and conversion. The safe temperature operating range was deter-
mined to be below 60 °C to avoid potential critical oxidation and
explosions due to periodic acid.[3] From all the reaction mixtures
submitted for both dynamic and isothermal DSC analysis, an on-
set of decomposition was typically observed at around 100 °C.
The measurements showed that within 30 minutes at a tempera-
ture above 120 °C, a highly energetic decomposition was taking
place (Fig. 1).[4]
Keywords: Androstadiene · E-factor · Environmental impact ·
Process Mass Intensity · Reaction Mass Intensity · Telescope
Introduction
Inhaled corticosteroids (ICS) are clinically proven to be some
of the most effective treatment for all levels of severity of asth-
ma. In the course of our program on ICS, we became interest-
ed in derivatives of the readily available epoxyparamethasone
2.[1] Several candidates with substituents on the hydroxyl and
ester anchors had shown remarkable potency. The synthetic ap-
proach of these derivatives was based on oxidative cleavage of
the α-hydroxyketone to the corresponding acid.[2] The resulting
acid was esterified and the free tertiary hydroxyl group was func-
tionalized with diverse heterocyclic moieties. The synthesis was
completed by acidic opening of the epoxide ring in stereoselec-
tive fashion to the α-chloro derivatives. A model substrate 1 is
described in the rest of the discussion to illustrate our newly de-
veloped process (Scheme 1), whereby the methyl ester resulting
from cleavage of the α-hydroxy ketone and the 3-methylthienyl-
A subsequent solvent screening revealed that only ethereal
solvents such as THF or 2-methyltetrahydrofuran (2-MeTHF)
could promote complete cleavage. All other solvents (aromatics,
alkanes, chlorinated or highly polar solvents, water and phase-
transfer catalysts) showed only partial or unselective reactivity
within the safe operating range.[5]
The reported chemical incompatibility of periodic acid with most
organic solvents[6] prompted us to reduce the concentration of
oxidizing agent in the organic phase by rendering the system bi-
phasic through the addition of water. In the case of water-misci-
ble organic solvents, an alkane co-solvent (ca. 20% depending
on the solvent) was added to obtain a heterogeneous system. In
all solvents except methanol, an incomplete conversion was ob-
served. Recourse to methanol was however not an alternative as
explosive methylperiodate could be formed.[7] The use of a phase
transfer catalyst was also not regarded as desirable as explosions
have been reported.[8]
At the outcome of the screening, a mixture of THF/hexanes/water
was shown to give the optimal results with complete conversion
in about 6 h. Interestingly, only marginal differences were ob-
served when switching to the simpler 2-MeTHF/water solvent
Scheme 1. Acidic opening of the epoxide ring in stereoselective fashion
to the α-chloro derivatives.
Scheme 2. Selective oxidative cleavage of the α-hydroxy ketone.

878 CHIMIA 2011, 65, No. 11
Columns
R2
R1
O
O
O
O
R1
R2
R1
R1
R1
R2
R2
OOH
R2
OO
O
O
O
O
or
O
R1 = Me, R2= H
or R1 = H, R2 = Me
R1 = Me, R2= H
or R1 = H, R2 = Me
Fig. 2. Oxidative cleavage. Left: water-soluble hydroperoxide; right:
water-insoluble ether peroxide.
Table 2. Peroxide content in the organic phase
Entry Amount
periodic
Reaction Reac- Content of Content of per-
tempe-
rature
tion
time
peroxide in oxide in the or-
the organic ganic phase af-
acid
phase
ter phosphinic
acid washing
1
2
3
4
5
1.1 equiv. 23 °C
1.1 equiv. 35 °C
1.3 equiv. 23 °C
1.3 equiv. 35 °C
1.3 equiv. 23 °C
16 h
16 h
16 h
16 h
6 h
101 ppm
99 ppm
88 ppm
106 ppm
66 ppm
< 5 ppm
Reaction conditions: 0.13 g/mL of 2 in 2-MeTHF
The reactive system had been designed on purpose as a bipha-
sic system in order to get a better phase separation and a lower
content of periodic acid in the organic phase. After complete con-
version, the excess of periodic acid and water-soluble peroxides
were removed from the organic phase by aqueous washings and
the peroxide content in the organic phase was measured (Table
2).[10]
Figure 1
Fig. 1. Representative isothermal and dynamic differential scanning
calorimetry curves.
In all cases, independent of the temperature or the stoichiometry
of periodic acid, a content of about 100 ppm was observed. Only
in the case of significantly reduced reaction time (entry 5, 6 h
reaction time instead of 16 h), the content was lowered margin-
ally (from ca. 100 ppm to 66 ppm), presumably due to reduced
peroxidation of the solvent itself. Eventually, the remaining per-
oxide could be removed by washing the organic layer with a 50%
aqueous solution of phosphinic acid thereby reducing the perox-
ide content to less than 5 ppm. The use of aqueous solutions of
sodium thiosulfate or sodium sulfite was not sufficient to reach
the targeted levels of peroxides of less than 5 ppm. Overall, the
desired oxidative cleavage product could be obtained in 94% as-
say yield and purity higher than 95%. The high purity at this stage
and the high potency of the intermediate prompted us to carry on
the next step without isolation and further purification.
system, which did not require the presence of an additional al-
kane solvent. The reaction rate was however reduced by a fac-
tor of almost 3 (ca. 15 h for completion).[8] This solvent system
nevertheless sparked our interest as it would lead to multiple op-
portunities for the downstream process.
The difficulty to remove unreacted starting material prompted
us to aim at full conversion in the oxidative cleavage. With the
MeTHF/water solvent system, the excess of oxidizing agent could
not be reduced to less than 1.1 equiv. as only incomplete conver-
sion or poorly reproducible results were obtained in reasonable
reaction times (less than 15 h). A stoichiometry of periodic acid
of 1.3 equiv. relative to the epoxyparamethasone starting material
was therefore used for the further optimization (Table 1).
From a safety standpoint, the transformation was obviously of
particular concern as it was associated with a high risk of perox-
ide formation when carried out in an ethereal solvent.
Esterification Steps
The subsequent esterification of the carboxylic acid required an-
hydrous conditions. While classical crystallization was a viable
option, direct use of the crude product in solution was envisioned
with an in situ drying process to minimize open handling of dried
product. The newly identified solvent system now allowed us to
proceed very efficiently via cycles of azeotropic distillations.[11]
This drying procedure was used effectively on multi-kg scale on
one of our targeted derivatives in the pilot plant. Once again,
2-MeTHF had turned out to be a superior alternative to THF due
to the improved efficiency of the azeotropic distillation.[12] The
longer reaction times in 2-MeTHF were partially compensated
by the reduced number of cycles and faster azeotropic drying. In
addition, the process allowed for easier recycling of the solvent
and minimization of the contaminated solvent waste.
2-MeTHF is known to form two different classes of peroxides
(Fig. 2). The first class, water-soluble, was regarded as more
easily manageable. It could indeed be controlled by removal via
aqueous washings. The second class, water-insoluble peroxides,
is actually less well-documented and presented more safety con-
cerns. Little is known about them apart from their qualitative
solubilities in organic solvents.[9]
Table 1. Optimization of a periodic acid stoichiometry relative to epoxy-
paramethasone
Stoichiometry periodic acid
0.95 equiv.
Results
incomplete conversion after 2 days
incomplete conversion after 2 days
complete conversion after 24 h
complete conversion after 12 h
1.0 equiv.
For the esterification of the free tertiary hydroxyl group, the orig-
inal process reported the use of toluene and THF as solvents for
the formation of the activated acid chloride 4 and the subsequent
1.1 equiv.
1.3 equiv.

Columns
CHIMIA 2011, 65, No. 11 879
Scheme 3.
S
HO
(2 eq)
O
SOCl2
2-MeTHF
S
O
O
O
O
HO
S
OH
O
S
O
HO
S
4
O
O
Cl
O
O
O
O
O
dbu
H2O
Pyridine
O
O
O
2-MeTHF
F
F
F
6
3
5
the desired product 7 in ca. 88% yield and purity >97%. While
the original isolation approach allowed the storage of a stable in-
termediate, it did not present any further advantages as the purity
was neither increased, nor was any specific impurity apparently
removed.
At this stage, a fully telescoped process using a single organic
solvent had been demonstrated without compromising on the
purity of the penultimate intermediate. The last chemical trans-
formation, namely the opening of the epoxide, turned out to be a
lot more problematic and challenging.
Scheme 4.
Ring-opening
Formation of the 6α-chloro side-product 1’ indeed proved to
be a major hurdle (Scheme 5). Its formation that resulted from
displacement of the allylic 6α-fluoride with retention of con-
figuration made it an inherent process-related impurity. No
6β-haloisomer was observed as expected due to the poor thermo-
dynamic stability of this isomer. It was also shown that it could
not be removed by recrystallization from levels higher than 3%,
and therefore had to be limited to below 3% in the crude reac-
tion mixture to guarantee its complete removal required by the
specifications at the Drug Substance stage.
Multiple hydrochloric acid sources were therefore evaluated to
minimize its formation. At best, levels around 5% of the unde-
sired impurity 1’ were observed when hydrogen chloride was
directly added as gas to a solution of epoxide in toluene at room
temperature. These levels could go up to 15% when the tempera-
ture raised even slightly. The amount of the 6α-chloro side-prod-
uct 1’ could eventually be significantly reduced when hydrogen
chloride was introduced in large excess into the reaction mixture
as an isopropanolic solution (Table 3). With this simple process
change, the amount of the 6α-chloro side-product 1’ could be
decreased after extensive optimization down to 0.7% in the crude
reaction mixture when a large excess of HCl was used (10 equiv.)
under kinetic control (0–5 °C, 15 h).
esterification to 6 (Scheme 3). This was rapidly streamlined
and adjusted to better align with our overall process strategy by
converting the 3-methylthienyl-2-carboxylic acid into the cor-
responding acid chloride 4 with thionyl chloride in 2-MeTHF
and using it directly in the esterification step in the same sol-
vent. Product formation proceeded via formation of the mixed
anhydride intermediate 5 that required at least 2 equiv. 3-me-
thylthienyl-2-carboxylic chloride 4 for complete conversion.
Stoichiometric quantities of this material 4 led to much slower
and incomplete reactions as esterification of the sterically hin-
dered tertiary alcohol competed with anhydride formation on the
acid moiety present in the molecule.
Compound 5 was obtained in high purity, even without purifica-
tion. It was therefore decided that the product would be used as
such in the next step. The esterification of the free acid 6 was
shown to proceed best using 1,8-DBU as a base in 2-MeTHF
(Scheme 4). In the case of the more challenging methyl ester for-
mation, dimethyl sulfate was used as the alkylating agent. After
the reaction had gone to completion, the excess of dimethylsulfate
was reacted with a large excess of morpholine to reduce the total
residual amount of alkylating agents to below 5 ppm. Compound
7 was originally crystallized from methanol and water. This pu-
rification required multiple operations as a solvent switch was
necessary, as well as handling of the dried powder. The latest
process improvements could however lead to comparable quality
of material at that stage in a fully telescoped approach to provide
Under these optimized conditions, the crude reaction mixture
could be directly crystallized from this same 2-MeTHF/iPrOH
solvent system to obtain the desired end-product as a transient
isopropanol solvate in 65% isolated yield over the whole five step
Scheme 5. Structure of 6α-chloro side-product
1’.

880 CHIMIA 2011, 65, No. 11
Columns
Table 3. Optimization results
Although the kinetic of the first oxidative cleavage step slightly
increased the individual reaction time, the overall cycle time
eventually improved dramatically due to a much reduced number
of operations and streamlined work-up protocols. In addition, the
new process increased the overall yield from 52% to 65%, mostly
due to reduced mechanical losses (loss of material in the workup
and purification steps in the original synthesis, and during the
isolation and purification operations). It is worth pointing out
that neither Process A nor Process B has really been optimized at
this stage. Finally the streamlined synthesis minimized the need
to handle dry and toxic material. Interestingly, the primary ben-
efits of the direct isolation process also consisted in the decreased
processing costs as very well described by Anderson in an earlier
review on the benefits of direct isolation processes.[14]
Entry HCl in i-PrOH / equiv. Yield
Product 1a
6α-Chloro
derivative 1’a
1
2
3
6.6
74%
74%
78%
94.8%
97.1%
97.7%
1.3%
0.9%
0.7%
9.9
13.2
aHPLC area% at 210 nm
sequence and 98% purity, similar to the results obtained with the
original process.
Discussion
The suitable physical properties of 2-Me-THF, more specifi-
cally its polarity, water-immiscibility and azeotropic properties
made this change possible. Although slightly more expensive
than THF, it eventually turned to be the solvent of choice for our
process both from an economic and environmental standpoint.
To illustrate these ‘first impressions’, we wanted to measure
some of the well-accepted green metrics such as the E-factor,
the Reaction Mass Efficiency, and the Process Mass Intensity.
Sheldon’s E-factor[15] (Em) is defined as the mass of total waste
per mass of product. The Reaction Mass Efficiency[16] (RME) di-
rectly related to the E-factor is defined as 1/(1+Em). The Process
Mass Intensity[17] (PMI) is defined as the quantity of raw mate-
rial input in kilogram per the quantity of bulk API in kilogram.
For the latter, several PMIs were calculated on solvents, water,
reagents.[13] The first generation of the synthesis was calculated
as a convergent sequence due to the formation of the acid chloride
that was done in a separate step. The second generation of the
synthesis was calculated as a single sequence, as no intermedi-
ate was isolated and only assay yields were available. Only the
overall yield and the overall waste production were used.
Throughout the development of this synthesis, a conscious effort
was made to find practical ways of improving the safety and envi-
ronmental impact of the process. Our efforts span over a range of
concerns such as the identification of the most efficient synthesis
with regard to atom economy and reaction yields, the use of safe
and less hazardous chemicals, the elimination and/or reduction of
waste as well as the number of operations with the additional goal
of reduced exposure to toxic, dry materials. These basic prin-
ciples, the foundation of green chemistry, are well-known to the
scientific community.[13] However, practical examples that illus-
trate their relevance are still scarce. We therefore wanted to dem-
onstrate quantitatively the relevance of some of the well-accepted
green chemistry metrics. The two latest processes will be used for
the comparison and discussion, namely the original process with
standard isolation of intermediates, referred to as Process A, and
the fully streamlined one described above, referred to as Process
B (Scheme 6).
As a result of this work, it had proven possible to replace three
organic solvents in the original synthesis, namely THF, hexane
and toluene with 2-MeTHF. All the intermediates were used in
the subsequent steps without purification in the original reaction
solvent, 2-MeTHF. This apparently simple change was success-
ful thanks to extensive screening as well as process understand-
ing and greatly impacted the overall productivity of the process.
The process is defined as all steps in the synthetic path starting
from the readily available epoxyparamethasone. Raw materials
are defined as all in-process materials that are used directly in the
chemistry of synthesizing, isolating, and purifying the bulk API.
The bulk API is the final form of the active ingredient that is pro-
duced in the synthesis, dried to the expected specifications, prior
Scheme 6. Process A: a) H5IO6, THF/hexane/
water, rt, 94%; b) SOCl2, toluene, 0–5 oC,
S
quant.; c) pyridine, THF, 0–5 oC, then d) DBU,
THF, 0–5 oC, and H2O, 84%; e) dimethyl sulfate,
THF, 0–5 oC, 96%; f) HCl, toluene, 0–5 oC, 68%.
Process B: a) H5IO6, 2-MeTHF / water, rt, not
HO
O
b
isolated; b) SOCl2, 2-MeTHF, rt, not isolated; c)
pyridine, 2-MeTHF, not isolated, then d) DBU,
2-MeTHF, 0–5 oC, and H2O, not isolated; e) di-
methyl sulfate, 2-MeTHF, 0–5 oC, not isolated;
f) HCl in iPrOH, 2-MeTHF, 0–5 oC, 65% overall
for the six chemical transformations.
S
HO
S
4
OH
OH
Cl
O
O
O
O
S
O
O
OH
O
O
O
O
O
a
c
O
O
O
F
F
F
5
3
2
O
O
O
O
S
S
S
HO
O
O
O
O
O
O
HO
O
O
O
d
e
f
Cl
F
O
O
O
F
F
1
6
7

Columns
CHIMIA 2011, 65, No. 11 881
Table 5. Costs comparison between the two processes
Table 4. Comparison of process A and process B
Process A
Process B
Process A
94%
Process B
n.i.
COG (USD)
/ kg end-
product
Cost per
step
(USD)
COG
Cost per
step
(USD)
Yield Step 1 (oxidative cleavage)
Yield Step 2 and 3 (esterification)
Yield Step 4 (esterification)
Yield Step 5 (ring opening)
PMI Step 1
(USD) /
kg end-
product
84%
96%
68%
51.8
6.4
44.3
403.1
9
n.i.
n.i.
65%
n.i.
Step 1
Step 2 and 3
Step 4
Step 5
Overall
3.9
10.7
5.3
13.9
44.4
129.5
355.6
181.8
674.1
1340.9
–
–
–
–
172.8
429.6
21.1
PMI Step 2 and 3
n.i.
26.6
PMI Step 4
n.i.
13.4
650.1
PMI Step 5
92.9
n.i.
COG: Cost of Goods
PMI solvents Step 1
PMI solvents Step 2 and 3
PMI solvents Step 4
35
n.i.
Conclusion
31
n.i.
The development of a streamlined process for our androstadiene
C-17 ester candidates led us to identify a fully streamlined pro-
cess that utilized only two solvents over five steps and a single
isolation of the end product. The overall amount of waste in ki-
logram per kilogram of final product, could be dramatically re-
duced. The improved ecological impact was also in accordance
with a reduced overall cost. Finally, the new process allowed for
much improved personal safety as it removed the need for isola-
tion and handling of highly active intermediates. We hope that
this successful example of the application of the principles of
Green Chemistry will be regarded as a useful example by the
scientific community and that the concept will be further used
particularly in the context of highly active compounds.
PMI solvents Step 5
401
476
41
n.i.
PMI solvent overall
52.8
n.i.
PMI water Step 1
PMI water Step 2 and 3
PMI water Step 4
49
n.i.
58
n.i.
PMI water Step 5
63
n.i.
PMI water overall
211
2
31.1
n.i.
PMI substrates and reagents Step 1
PMI substrates and reagents Step 2 and 3
PMI substrates and reagents Step 4
PMI substrates and reagents Step 5
PMI substrates and reagents overall
PMI overall
6
n.i.
11
n.i.
15
n.i.
34
9
Experimental Section
515.0
92.9
50.7
PMI overall predicted with 80% solvent 515.0
recycling
(6α,9β,11β,16α,17α)-9,11-epoxy-6-fluoro-17-hydroxy-
16-methyl-3-oxo-Androsta-1,4-diene-17-carboxylic
acid (3)
Em
512.7
0.195%
58%
81.6
RME
1.21%
65%
Epoxyparamethasone (2) (99.6 g, 255 mmol, 1.0 equiv.) was
suspended in 2-MeTHF (0.8 L) and a freshly prepared aqueous
solution of periodic acid (81.4 g, 357 mmol, 1.4 eq into 0.22 L
water) was added at a rate such that the temperature remained
below 30 °C (ca. 15 min addition). After the conversion is com-
plete (about 12 h), the phases were separated, and the organic
phase was washed twice with 0.25 L water. The water washings
were combined and kept separately. The organic phase was then
washed twice with an aqueous phosphinic acid solution (33 mL
of the commercial 50% phosphinic acid solution in water, further
diluted with 0.23 L water). It is critical that these washings are
kept separately and not combined with the previous water
extracts as phosphinic acid would react violently with the ex-
cess periodic acid. At this point, no peroxide was observed in
the organic phase. The crude solution was then concentrated to
ca. half the volume under reduced pressure (ca. 300–400 mbar,
internal temperature ca. 50–55 °C). An equal amount of dry
2-MeTHF (ca. 0.5 L) was added and the azeotropic distillation
was repeated twice more to result in a dry crude solution of prod-
uct in 2-MeTHF (Assay yield: 94%, purity: ca. 96% by HPLC).
1H NMR (DMSO-d6, 400 MHz) 0.84 (d, J=7.2 Hz, 3H), 0.88 (s,
3H), 1.18–1.24 (m, 1H), 1.36 (s, 3H), 1.38–1.72 (m, 3H), 2.18 (d,
J=13.5 Hz, 2H), 2.30 (ddd, J=10.8, 7.8, 6.5 Hz, 2H), 2.56–2.64
(m, 2H), 2.71–2.79 (m, 1H), 3.33 (br s, 1H), 5.62–5.78 (2m, 1H),
6.15 (d, J=2.0 Hz,1H), 6.18 (s, 1H), 6.60 (dd, J=9.8, 1.3 Hz, 1H).
Overall yield
n.i. not isolated; PMI: Process Mass Intensity; Em: E-factor; RME:
Reaction Mass Efficiency
to any physical modification steps such as milling or formulation.
Atom economy did, of course, not change as exactly the same re-
actions were used in the two versions of the process. The results,
shown in Table 4, clearly show the improvement in ecological
impact and reaction mass efficiency for the improved synthesis.
Both E and RME have improved by a factor of around 6. The
overallmamount of waste in kilogram per kilogram of final prod-
uct, could be dramatically reduced from 515 kg/kg to 93 kg/kg.
During the synthesis solvents and reagents were not recovered,
which could provide another point of improvement for the pro-
cess in the future. With 80% recycling efficiency in process B for
example, the overall PMI would be further reduced to ca. 51 kg/
kg. Recycling would not be realistic in process A.
A cost of goods analysis was also made to evaluate the evolu-
tion of the cost with the new process. Processing costs, which
would further reemphasize the benefit of direct isolation process
B, were omitted. The price calculation was made in US dollars
based on prices taken directly from the common vendors (Sigma
Aldrich). The streamlined process B turns to be more advanta-
geous, despite the omission of processing costs. While the first
two steps turned to be more expensive than in processA that used
cheaper solvents, it was significantly compensated by the more
efficient processes used in the subsequent steps. The minimized
mechanical losses also helped bringing down the overall cost ul-
timately by a factor 3 (Table 5).
(6α,9β,11β,16α,17α)-9,11-epoxy-6-fluoro-16-methyl-
17-[[(3-methyl-2-thienyl)carbonyl]oxy]-3-oxo-Andros-
ta-1,4-diene-17-carboxylic acid methyl ester (7)
Preparation of 3-methyl-2-thiophenecarbonyl chloride (4): 3-me-
thylthienyl-2-carboxyl acid (79 g, 0.56 mol, 2.1 equiv.) was dis-

882 CHIMIA 2011, 65, No. 11
Columns
L. Mazzoni, A. Nicholls, G. E. Pilgrim, E. Shaebulin, G. M. Spooner, R.
Stringer, P. Tranter, K. L. Turner, M. F. Tweed, C. Walker, S. J. Watson, B.
M. Cuenoud, Bioorg. Med. Chem. 2004, 12, 5213; Novartis Pharma AG,
B. Cuenoud, D. Beattie, T. H. Keller, G. E. Pilgrim, D. A. Sandham, S. J.
Watson, WO2002/00679 A2; Novartis Pharma AG, G. Jordine, M. Mutz,
WO2004/013156 A1.
solved into 2-MeTHF (0.3 L) and cooled to 0 °C. Thionyl chlo-
ride (151 g, 92 mL, 1.27 mol, 2.4 equiv.) was added at rate such
that the temperature remained below 5 °C (ca. 15 min addition)
and the resulting mixture was stirred at that temperature for 1 h
and used as such in the subsequent transformation.
[2] D. J. Kertesz, M. Marx, J. Org. Chem. 1986, 51, 2315.
To the freshly prepared solution of 3 (ca. 100 g, ca. 0.27 mol) in
2-MeTHF (0.45 L) was added dry pyridine (147.5 g, 7.3 equiv.),
and the freshly prepared solution of 3-methyl-2-thienyl carboxyl
chloride 4 (freshly prepared solution from above) was added at
a rate such that the temperature remained below 5 °C (ca. 30
min addition). The mixture was stirred for an additional 2 h at
0–5 °C. Water (ca. 0.4 L) and dbu (102 g, 670 mmol, 2.5 equiv.)
were added in ca. 15 min at that temperature and the resulting
mixture was stirred at that temperature for ca. 2 h, until complete
hydrolysis of the mixed anhydride. The pH was adjusted to 1–1.5
by addition of hydrochloric acid, and the aqueous layer was re-
moved. Dimethyl sulfate (95 g, 750 mmol, 2.8 equiv.) was added
at that temperature and the resulting mixture was stirred for 15 h
until complete methyl ester formation, quenched with the addi-
tion of morpholine (58 g, 665 mmol, 2.5 equiv.) and stirred for an
additional 30 min. The reaction mixture was quenched with water
(0.2 L) and the aqueous layer was removed. The organic layer
was washed with water (0.2 L), dilute HCl (2M, 0.2 L), and brine
(0.2 L). The resulting organic extract was concentrated to ca. half
the volume under reduced pressure to afford the desired product 7
in ca. 88% assay yield and ca. 97% purity. 1H NMR (DMSO-d6,
400 MHz) 0.90 (d, J=7.2 Hz, 3H), 0.91 (s, 3H), 1.20–1.26 (m,
1H), 1.39 (s, 3H), 1.50–1.93 (m, 2H), 2.20–2.32 (m, 1H), 2.41 (s,
3H), 2.46–2.75 (m, 3H), 3.16–3.37 (m, 1H), 3.57–3.84 (m, 1H),
3.65 (s, 1H), 3.78 (s, 3H), 5.63–5.79 (m, 1H), 6.19 (dd, J=1.8,
10.1 Hz, 1H), 6.25 (s, 1H), 6.57–6.65 (m, 1H), 7.05 (d, J=5.0 Hz,
1H), 7.13–7.25 (m, 1H), 7.74 (t, J=6.1 Hz, 1H).
[3] L. Pacesova, Z. Hauptman, Z. Anorg. Allg. Chem. 1963, 325, 325; O. N.
Breusov, N. I. Kashina, T. V. Revzina, Zh. Neorg. Khim. 1970, 15, 612.
[4] J. Schmidlin, Chem. Abstract 1985, 103, 105215.
[5] ‘Periodic Acid’, ‘Encyclopedia of Reagents for Organic Synthesis’, Wiley
& Sons, Posted September 15, 2006. http://onlinelibrary.wiley.com/o/eros/
articles/rn00561/pdf_fs.html
[6] Methyl periodate is a reported explosive. See B. Martel, ‘Chemical risk
analysis: a practical handbook’, Hermes Penton, London, UK, 2004, p. 280.
[7] T. Takata, R. Tajima, W.Ando, J. Org. Chem. 1983, 48, 4764; P. Ferraboschi,
M. N. Azadani, E. Santaniello, S. Trave, Synth. Commun. 1986, 16, 43.
[8] F. D. Albinson, S. J. Coote, J. M. Robinson, J. Malcolm, PCT Int. Appl.
2002, WO 200200843 A1 20020131.
[9] K. V. S. Rao, K. V. S. Rama, A. V. Sapre, Radiation Effects 1970, 3, 183;
V. G. Glukhovtsev, G. I. Zaitseva, Y. V. Il’in, J. Org. Chem. USSR 1981,
17, 438; V. G. Glukhovtsev, Y. V. Il’in, G. I. Nikhishin, Bull. Acad. Sci.
USSR Div. Chem. Sci. 1983, 32, 1479; Y. V. Il’in, V. A. Platkhotnik, V.
G. Glukhovtsev, G. I. Nikhishin, Bull. Acad. Sci. USSR Div. Chem. Sci.
1982, 31, 1976; E. M. Kuramshin, V. K. Gumerova, S. S. Zlot-skii, D. F.
Rakhmankulov, V. G. Glukhovtsev, J. Gen. Chem. USSR 1983, 53, 148; V.
V. Zorin, S. S. Zlot-skii, V. G. Glukhovtsev, Y. V. Il’in, G. I. Nikhishin, D.
F. Rakhmankulov, J. Org. Chem. USSR 1981, 17, 231; K. Griesbaum, G. H.
Mertens, P. Krieger-Beck, H. Henke, Can. J. Chem. 1994, 72, 2198; G. I.
Nikhishin, et al. Bull. Acad. Sci. USSR Div. Chem. Sci. 1971, 20, 2202.
[10] S. Baj, Fresenius J. Anal. Chem. 1994, 350, 159.
[11] The new solvent system 2-MeTHF/water allowed us to start the drying
with a water content twice as low as with the former solvent system THF/
heptane/water (ca. 5% residual water instead of >10%). Besides, the
azeotropic drying proceeded within 2 cycles instead of 4 or 5 typically
required in the presence of THF, in the absence of a rectification column.
[12] Boiling point azeotrope 71 °C at atmospheric pressure, 89.4 wt% solvent,
10.6 wt% water.
[13] P. Anastas, J. Warner, ‘Green Chemistry: Theory and Practice’, Oxford Uni-
versity Press, Oxford, UK, 1998, p. 30.
[14] N. G. Anderson, Org. Proc. Res. Dev. 2004, 8, 260.
[15] R. A. Sheldon, Chem. Ind. (London), 1992, 903.
[16] J. Andraos, Org. Process. Res. Dev. 2005, 9, 149; J. Andraos, Org. Process.
Res. Dev. 2005, 9, 404.
[17] C. Jimenez-Gonzalez, C. S. Ponder, Q. B. Broxterman, J. B. Manley, Org.
Process. Res. Dev. 2011, 15, 912.
(6α,11β,16α,17α)-9-chloro-6-fluoro-11-hydroxy-16-
methyl-17-[[(3-methyl-2-thienyl)carbonyl]oxy]-3-oxo-
Androsta-1,4-diene-17-carboxylic acid, methyl ester
(1)
To a solution of compound 7 (crude mixture from above) in
2-MeTHF (ca. 0.5 L) was added a 6.8 M solution of hydrogen
chloride in i-PrOH (0.5 L, 3.4 mol, 13.2 equiv.) at 0–5 °C in 30
min. The solution was stirred for 15 h until completion. The mix-
ture was concentrated to ca. half volume under reduced pressure
o
(ca. 100 mbar, internal temperature ca. 55 C) and the process
was repeated after addition of ca. 0.2 L of i-PrOH. An additional
0.1 L of i-PrOH was added and the mixture was heated to ca. 60
oC, and seeded with the pure i-PrOH solvate. The mixture was
aged for 1 h, and cooled linearly to 0 °C in 6 h. The precipitate
was filtered off and washed with cold i-PrOH (0.1 L). Drying
at 40 °C afforded ca. 90 g of the pure i-PrOH solvate 1 as a
white powder in 78% yield and 97% purity. 1H NMR (DMSO-d6,
400 MHz) 0.90 (d, J=7.2 Hz, 3H), 1.00 (s, 3H), 1.20–1.26 (m,
1H), 1.61 (s, 3H), 1.62–1.72 (m, 1H), 1.85–1.93 (m, 1H), 2.23–
2.27 (m, 1H), 2.43 (s, 3H), 2.46–2.55 (m, 3H), 2.75 (dt, J=3,0,
8.0 Hz, 1H), 3.25–3.31 (m, 1H), 3.65 (s, 3H), 4.39 (br s, 1H),
5.53–5.71 (m, 1H), 5.65 (d, J=11.5 Hz, 1H), 6.11 (s, 1H), 6.29
(dd, J=2.0,10.3 Hz, 1H), 7.04 (d, J=5.0Hz, 1H), 7.28 (dd, J=1.5,
10.1Hz, 1H), 7.75 (d, J=4.7 Hz, 1H); 13C NMR (DMSO-d6, 100
MHz) 15.5, 16.2, 16.9, 24.1, 32.6, 33.8, 33.9, 35.9, 36.0, 40.1,
48.7, 49.8, 51.8, 73.7, 83.8, 85.8, 87.5, 91.8, 119.5, 124.9, 128.7,
132.2, 146.6, 151.9, 160.9, 162.6, 169.1, 184.4; HRMS (M+H)
calcd 551.16700, obsd 551.16636.
Received: October 2, 2011
[1] D. A. Sandham, L. Barker, D. Beattie, D. Beer, L. Bidlake, D. Bentley, K.
D. Butler, S. Craig, D. Farr, C. Ffoulkes-Jones, J. R. Fozard, S. Haberthuer,
C. Howes, D. Hynx, S. Jeffers, T. H. Keller, P. A. Kirkham, J. C. Maas,