An efficient large-scale synthesis of a naphthylacetic acid CRTH2 receptor antagonist

Article abstract of DOI:10.1021/op300306c

An efficient and practical synthesis of a naphthylacetic acid CRTH2 receptor antagonist is reported. Michael addition of ethyl t-butyl malonate to an allenoate afforded a triester, which was selectively hydrolyzed and decarboxylated to give a benzylidenepentanedioic acid monoester. Treatment of this compound with potassium acetate and acetic anhydride produced the naphthylacetate core. The triflate of the key building block was coupled with a zinc reagent of the side chain under improved Negishi coupling conditions to afford the target product. The process was successfully scaled up to produce over 2 kg of the API.

Full text of DOI:10.1021/op300306c

Article
pubs.acs.org/OPRD
An Efficient Large-Scale Synthesis of a Naphthylacetic Acid CRTH2
Receptor Antagonist
Lianhe Shu,* Ping Wang, Chen Gu, Wen Liu, Lady Mae Alabanza, and Yingchao Zhang
Process Research and Synthesis, Pharma Early Research and Development (pRED), Hoffmann-La Roche, Inc., 340 Kingsland Street,
Nutley, New Jersey 07110, United States
S
* Supporting Information
ABSTRACT: An efficient and practical synthesis of a naphthylacetic acid CRTH2 receptor antagonist is reported. Michael
addition of ethyl t-butyl malonate to an allenoate afforded a triester, which was selectively hydrolyzed and decarboxylated to give
a benzylidenepentanedioic acid monoester. Treatment of this compound with potassium acetate and acetic anhydride produced
the naphthylacetate core. The triflate of the key building block was coupled with a zinc reagent of the side chain under improved
Negishi coupling conditions to afford the target product. The process was successfully scaled up to produce over 2 kg of the API.
it was decided to abandon this approach and explore alternative
syntheses.
Retro-synthetic analysis revealed several potential approaches
to 18. Of these, three disconnection strategies (Scheme 2) were
selected for investigation.
Route (a) was initially considered the most promising since a
highly selective annulation process of a related compound has
been previously described in the literature.4 The synthesis of
the precursor of 16, however, was found to be a challenge.
Though the Heck reaction of a close analogue of 17 is known
in the literature,5 no reaction between 17 and 23 was detected
under similar conditions (Scheme 3). Friedel−Crafts reactions
of analogues of 25, an alternative approach, have also been
reported.6 However, with compound 25, the reactions were
poor under all conditions examined (Scheme 3). The lack of
reactivity for 25 is likely due to the deactivating effect of
fluorine.
INTRODUCTION
■
CRTH2, chemoattractant receptor-homologous molecule ex-
pressed on T-helper-type cells, is one of the prostaglandin D2
receptors that is expressed on effector cells involved in allergic
inflammation such as T-helper type 2 (Th2) cells, eosinophils,
and basophils.1 Studies suggest that CRTH2 plays a
proinflammatory role in inflammation diseases. Thus, a
CRTH2 receptor antagonist could be therapeutically useful
for the treatment of asthma, allergic inflammation, COPD,
allergic rhinitis, and atopic dermatitis.2 Over the past few years,
a number of distinct CRTH2 antagonists have been reported
that gave encouraging proof-of-concept results in clinical trials.2
At Roche, compound 14, a naphthylacetic acid, was identified
as a potent CRTH2 receptor antagonist with good pharmaco-
logic properties. In order to support further preclinical/clinical
development, process research was initiated to develop a
practical synthesis of this molecule in bulk quantities.
The original synthesis of 14 (Scheme 1)3 was not suitable for
large-scale production due to several issues: (1) chromato-
graphic purification were required for seven steps; (2)
environmentally unfriendly or safety hazardous reagents, such
as carbon tetrachloride and lithium aluminum hydride were
used; (3) the formation of the zincate 12 for the Negishi
coupling step was highly exothermic under the original
conditions; and (4) the high catalyst loading in the penultimate
step for the Negishi coupling could result in problems with
palladium removal. Herein, we describe the development of a
practical synthesis of 14, which has been successfully scaled up
to produce >2 kg of active pharmaceutical ingredient (API).
The preparation of precursors for route (b) was also
problematic. The coupling of 26 with 27 was studied under a
variety of conditions described in the literature.7 Unfortunately,
none afforded a clean reaction that had the potential for scale-
up (Scheme 4). This result was somewhat consistent with the
observation of Davies et al.,7a that reactions of methyl
propiolate with 2-fluorobenzyl chloride or 4-trifluoromethyl-
benzyl chloride gave complex mixtures.
On the other hand, route (c) intermediate allenoate 32 was
easily prepared via reaction of 31 with acid chloride 30 under
typical Wittig reaction conditions.8 The subsequent Michael
addition with a Meldrum’s acid proceeded cleanly to give 34 in
97% yield. With 34 in hand, we envisioned that an
intramolecular Friedel−Crafts reaction would directly provide
35 with elimination of acetone and carbon dioxide (Scheme 5).
Though catalytic intramolecular Friedel−Crafts reactions of
Meldrum’s acids have been utilized for the synthesis of 1-
tetralones,9 compound 34 was found to be stable under neutral
or weak acidic (e.g., boric acid) conditions; no cyclization was
RESULTS AND DISCUSSION
■
Synthesis of the Naphthylacetate Core. In the original
synthesis, the naphthylacetate 9 was prepared in seven steps
from 1. Five of the seven steps were, however, for the
homologation of 4, which was considered poor atom economy,
despite good yields obtained for each individual step. The most
problematic step was the condensation of 1 with 2, which was
poorly selective and produced many byproducts. As a result, the
product isolation was tedious, and the yield was low. Therefore,
Received: October 26, 2012
Published: March 18, 2013
© 2013 American Chemical Society
651
dx.doi.org/10.1021/op300306c | Org. Process Res. Dev. 2013, 17, 651−657

Organic Process Research & Development
Article
a
Scheme 1. Original synthesis of 14
a
Reagents and conditions: a) NaH, rt, chromatography, 33%; b) (i) (CF3CO)2O, TEA, (ii) NaBH4, MeOH, chromatography, 80%; c) BnBr, K2CO3,
chromatography, 94%; d) LiAlH4, 92%; e) PPh3, CCl4, chromatography, 83%; f) Pd(PPh3)2Cl2, CO, MeOH, chromatography, 97%; g) H2, 10% Pd/
C, 98%; h) Tf2O, pyridine, CH2Cl2, chromatography, 84%; i) Zn, LiCl, BrCH2Cl2Br, TMSCl, THF; j) Pd(OAc)2 (0.1 equiv), S-Phos (0.2 equiv),
chromatography, 98% from 10; k) aq LiOH, THF, 92%.
Scheme 2. Retro-synthetic analysis of naphthylacetate 18
Scheme 3. Attempted preparation of precursor 24
potassium acetate in acetic anhydride resulted in the formation
of a mixed anhydride intermediate which, upon stirring at 80
°C overnight, was converted to a new compound with a
molecular weight matching that of 35. Attempts to isolate this
major component in the pure form were not successful as it
Scheme 4. Attempted preparation of 28
1
decomposed during the work-up. H NMR spectrum of the
crude product showed the coupling pattern of a para-
substituted aromatic ring, and LC−MS analysis did not give
the expected [M − 1] peak in the negative mode. The lack of
observed after stirring in xylene at reflux overnight. With
stronger acids, such as acetic acid, decarboxylation occurred,
and monoacid 36 was obtained. Treatment of 36 with
652
dx.doi.org/10.1021/op300306c | Org. Process Res. Dev. 2013, 17, 651−657

Organic Process Research & Development
Article
Scheme 5. Attempted preparation of 35 via 34
Scheme 6. Preparation of 35 via 41
Scheme 7. New synthesis of naphthylacetic acid core 50
stability combined with the spectral result suggested that a
competing cyclization occurred, and 37 was likely the major
product.
Initially, we presumed that the double bond in 36 would
migrate or isomerize under the reaction conditions. Therefore,
the configuration of this double bond would not be important.
The formation of 37 prompted us that a substrate with a Z-
double bond may be required to suppress this detrimental
cyclization pathway. Thus, the Z-isomer 41 was prepared in
several steps from 29 for evaluation. Although compound 39,
an analogue of 41, also underwent a similar cyclization pathway
as 36 to give 40, treatment of 41 with trifluoroacetic anhydride
and triethylamine indeed afforded the desired cyclization
product, presumably via a ketene intermediate. After hydrolysis
653
dx.doi.org/10.1021/op300306c | Org. Process Res. Dev. 2013, 17, 651−657

Organic Process Research & Development
Article
and aqueous work-up, a few milligrams of 35 was obtained after
purification by chromatography followed by crystallization
(Scheme 6), and its structure was confirmed by NMR analysis.
With the encouraging result of 41, the synthetic plan was
revised, as the preparation of 41 was poorly selective and low
yielding. Instead of 38, allenoate 43 was produced via the
Wittig reaction of 30 with phosphorane 42. Subsequent
Michael addition with tert-butyl ethyl malonate (44) afforded
triester 45, in which the ethyl esters could be selectively
hydrolyzed to give the tert-butyl ester 46. Upon heating to
reflux, 46 was decarboxylated to give the monoacid monoester
47 containing the desired Z-double bond (Scheme 7).
A solution of 30 in toluene and methyl tert-butyl ether
(MTBE) was added to a suspension of 42 and DIPEA in a
mixture of MTBE and n-heptane at around 5 °C. The solid
byproduct, triphenylphosphine oxide, was removed by
filtration, and 43 was obtained as a solution in >97% HPLC
purity after aqueous work-up. Compound 43 was not stable at
room temperature; HPLC analysis indicated formation of a new
broad peak at the level of ∼50% after 43 was stored as a neat oil
at room temperature for a week. LC−MS showed a molecular
weight of 440, suggesting a dimerization product, presumably
51 (Figure 1). In order to slow this dimerization process, 43
Monoacid 47 was added to a suspension of potassium acetate
in acetic anhydride and n-heptane. The mixture was warmed to
50 °C to induce the formation of mixed anhydride intermediate
48 (Scheme 7). The reaction temperature was then raised to 85
°C, and the cyclization was complete within 17 h. Upon
aqueous work-up, compound 49, the first crystalline
intermediate in the reaction sequence, precipitated from n-
heptane and was isolated in 99% HPLC purity and 62−69%
overall yield from 29. The acetyl group was then easily
hydrolyzed by treatment with aqueous potassium hydroxide.
The resulting naphthol 50 was obtained in 97% yield.
Since the hydrolysis of 45 was carried out in ethanol, the
Michael addition to 43 was examined in this solvent using
sodium ethoxide as the base. The resulting solution of 45 in
ethanol was treated with NaOH (2 equiv), and decarboxylated
to afford 47 as the major product. While this process eliminated
the isolation of 45, the downstream steps using this material,
however, gave 49 in lower yield and quality.10
Final Coupling and Hydrolysis. Triflate 53 was prepared
by the reaction of 50 with triflic anhydride using pyridine as the
base. After aqueous work-up, 53 was crystallized from DMF−
water and collected by filtration in 98% yield (Scheme 8).
The Negishi coupling of triflate 53 with 12 was very
problematic. In the original synthesis, this step was carried out
under typical conditions (Scheme 2).11 A mixture of predried
zinc dust and lithium chloride was dried under high vacuum at
170 °C and then suspended in anhydrous THF and treated
three times with 1,2-dibromoethane at reflux, followed by
further treatment with trimethylchlorosilane. This zinc
activation process is highly exothermic with vigorous foaming
that clearly cannot be run on a large scale. In the absence of
lithium chloride, the zinc activation was slightly less exothermic,
but the subsequent Negishi coupling was found to be
inconsistent. Since compound 11 is a benzyl chloride, which
should be relatively reactive to zinc, such a harsh zinc activation
process may not be necessary. A traditional zinc activation
method12 was thus investigated. Zinc dust was suspended in
DMF and treated with trimethylchlorosilane (0.1 equiv to 53).
This activation process was only mildly exothermic; with 237 g
of zinc, only a 5 °C temperature increase was observed over a
period of 30 min. Hydrogen chloride, which was generated
from trimethylchlorosilane and residual water in DMF, was the
actual reagent for the zinc surface cleaning. On scale-up, when
the water level in the DMF was lower, it was compensated by
the addition of isopropanol.
Figure 1. Dimerization product 51 and over-hydrolysis product 52.
was stored as an MTBE solution at 0 °C prior to further
processing. Under these storage conditions, the degradation
over one week was <1.5%.
In the Michael addition step, the solution of 43 in MTBE
was added to the enolate prepared by addition of 44 to a
suspension of sodium tert-butoxide in DMF. Triester 45 was
obtained as an oil in ∼95% HPLC purity and was directly used
in the next step.
The conversion of 45 to 47 was carried out in ethanol by
treatment with 3 equiv of lithium hydroxide. Once the
hydrolysis was complete, the reaction mixture was heated to
reflux to promote the decarboxylation. After careful aqueous
work-up, 47 was obtained as an oil in 92−93 area% HPLC
purity, which was directly used in the next step without
purification. The major byproduct 52 (Figure 1) was removed
by washing with aqueous sodium carbonate.
After activation of the zinc, a portion (20−50%) of a DMF
solution of 11 was added, and the temperature was raised to
30−45 °C. Once the zinc insertion was initiated, the remaining
Scheme 8. Improved Negishi coupling of 53
654
dx.doi.org/10.1021/op300306c | Org. Process Res. Dev. 2013, 17, 651−657

Organic Process Research & Development
Article
solution of 11 was added slowly to maintain the temperature at
∼45 °C. The reaction was typically complete within 1 h. An
excess of zinc (3 equiv to 53) was required to suppress the
formation of the homocoupling product 55 (Figure 2).
Additional development work would be required for further
scale-up.
EXPERIMENTAL SECTION
■
General. HPLC analyses were performed using an Agilent
1100 system with a HALO C8 (100 mm × 3 mm, 2.7 μm)
column, 30−100% CH3CN−water (+0.1% TFA) as mobile
phase at flow rate of 0.5 mL/min over 10 min, and UV
detection at 260 nm. HRMS (ESI, positive mode) was obtained
by an orbit-trap mass analyzer. Zinc dust (≤10 μm) was
purchased from Sigma-Aldrich and used as received.
Figure 2. Byproducts from the Negishi coupling step.
4-(4-Fluorophenyl)-2-methylbuta-2,3-dienoic Acid
Ethyl Ester (43). A reactor was sequentially charged with 4-
fluorophenylacetic acid 29 (8.24 kg, 53.4 mol), DMF (0.95 kg,
1.30 mol), and toluene (34 kg). While agitating at 21−26 °C,
oxalyl chloride (7.6 kg, 59.9 mol) was added via a metering
pump. The resulting mixture was agitated at 18−20 °C for 33
min, then concentrated under reduced pressure to ∼20 L. The
resulting concentrate was diluted with MTBE (16 kg) to give a
solution of 30 in MTBE−toluene.
The loadings of the palladium acetate and S-Phos were also
reduced to 0.5 and 1.0 mol %, respectively. However, even at
this level, the cost of the S-Phos was still considered too high
for large-scale production. Therefore, a readily available and
inexpensive palladium complex, bis(triphenylphosphino)palla-
dium chloride, was tested, and the reaction worked equally well
using a 0.5 mol % loading.
Due to safety concerns with transferring a zinc reagent from
one vessel to another in the lab setting, the coupling step was
carried out in one pot. Thus, to the zinc reagent, prepared as
described above, was added Pd(PPh3)2Cl2 (0.5 mmol %),
followed by 53 as a solid in one portion. Since no reaction
occurred below 50 °C, the reaction mixture was warmed to
∼60 °C. The coupling was initiated at this temperature and was
complete within 30 min to 1 h as indicated by HPLC analysis.
The reaction mixture was quenched with water and diluted with
isopropyl acetate. The resulting mixture was treated with N-
acetyl-L-cysteine to assist in removal of the palladium catalyst.
After crystallization from isopropyl acetate−n-heptane, com-
pound 54 was collected by filtration in 95% yield and >99%
purity as a white solid. The two major byproducts, 55 and 56
(Figure 2), were easily removed as they remained in the filtrate.
The reaction has been successfully and consistently carried
out multiple times on a 500-g scale. However, once the reaction
was initiated at 60−65 °C, a 30 °C temperature increase over
10 min was observed, while the external heating was removed.
Attempts to control the exotherm by a slow addition of 53 in
DMF to a preheated mixture of 12 in the presence of the
catalyst resulted in inconsistent reactions. In addition, once the
reaction stalled, it could not be resumed unless fresh zinc
reagent 12 was added, indicating that the zinc reagent may lose
reactivity in the reaction media. Therefore, for further scale-up,
adding a solution of 12 to the mixture of 53 and the catalyst
might be still required,13 or alternatively, the reaction could be
run in a continuous manner.14
A reactor was charged with (carboethoxyethylidene)-
triphenylphosphorane 42 (20.0 kg, 55.2 mol) and MTBE (45
kg). The resulting suspension was agitated at ∼20 °C for 12
h,15 and then DIPEA (8.5 kg, 65.8 mol) and MTBE (30 kg)
were added. After the mixture was cooled to 7 °C, the solution
of 30 from above was added over 2.3 h, and the empty reactor
was rinsed with MTBE (8 kg). The resulting suspension was
agitated at 5−6 °C for 30 min and then dropped into a
centrifuge. The collected solid was washed with MTBE−n-
heptane (1:1 mixture, three times, total 79 kg). The combined
filtrate and wash were charged back to the reactor. Then a
mixture of citric acid (50%, 15 kg) in water (28 kg) was
cautiously added over 41 min at ∼9 °C. The organic layer was
separated and washed with water (2 × 41 kg) and then
concentrated and diluted with MTBE (7.4 kg) to give a
1
solution of 43 in MTBE (∼40 kg, 53.4 mol in theory). H
NMR (300 MHz, DMSO-d6) 7.35 (m, 2H), 7.20 (m, 2H), 6.82
(q, J = 2.9 Hz, 1H), 4.14 (m, 2H), 1.92 (d, J = 2.9 Hz, 3H),
1.16 (t, J = 7.2 Hz, 3H); 13C NMR (100 MHz, DMSO-d6)16
211.35, 129.09, 129.01, 128.50, 116.01, 115.79, 98.68, 96.02,
60.78, 14.89, 14.12; HRMS (ESI) m/z calcd for C13H14O2F [M
+ H]+ 221.0978, found 221.0976.
2-Ethoxycarbonyl-3-[1-(4-fluorophenyl)meth-(E)-yli-
dene]-4-methylpentanedioic Acid 1-tert-Butyl Ester 5-
Ethyl Ester (45). A 400 L reactor was charged with NaOtBu
(5.34 kg, 55.6 mol) and DMF (77 kg). t-Butyl ethyl malonate
44 (11.5 kg, 61.1 mol) was added over 1 h at ∼20 °C, followed
by the solution of 43 (∼40 kg, 53.4 mol in theory) from above
over 100 min. The empty drum was rinsed with MTBE (7 kg),
which was added to the reactor. The resulting solution was
agitated at ∼20 °C for 70 min, after which HPLC analysis
indicated an acceptable level of 0.31 area % of 43. The reaction
mixture was then diluted with MTBE (45 kg). After the mixture
was cooled to 15 °C, AcOH (4.0 kg, 66.6 mol) was cautiously
added over 20 min at ∼15 °C, followed by water (66 L) over
40 min at 18 3 °C. The organic phase was separated, washed
with water (2 × 50 kg), and concentrated to dryness at 16−34
°C/5−100 mmHg vacuum (85.5 L of solvent was removed) to
give 45 (∼21.8 kg, 53.4 mol in theory) as an oil. The resulting
material was used directly in the next step without further
purification. 1H NMR analysis showed a mixture of triester and
its enol forms. HRMS (ESI) m/z calcd for C22H29FO6Na [M +
Na]+ 431.1846, found 431.1833.
The hydrolysis of 54 to 14 was carried out in acetic acid with
HCl at ∼45 °C. The reaction was complete within 30 min to
give a thick slurry. After aqueous work-up, 14 was isolated by
filtration in 95% yield and >99.5% purity as a white solid. The
residual metal level was well controlled: 6 ppm of palladium
and 20 ppm of zinc.
CONCLUSION
■
In summary, we have developed a new synthesis of a potent
CRTH2 antagonist in eight linear steps and 53% overall yield
from readily available 29. A novel telescoped process for the
preparation of naphthylacetate 49 from allenoate 43 has been
successfully carried out in the pilot-plant scale. Negishi coupling
between 53 and 12 under improved conditions was successfully
carried out at 500-g scale to produce >2 kg of the API.
655
dx.doi.org/10.1021/op300306c | Org. Process Res. Dev. 2013, 17, 651−657

Organic Process Research & Development
Article
3-[1-(4-Fluorophenyl)meth-(Z)-ylidene]-2-methylpen-
tanedioic Acid 5-tert-Butyl Ester (47). The 400 L reactor,
containing 45 (∼21.8 kg, 53.4 mol in theory) from above, was
charged with ethanol (88 kg). The mixture was cooled to 15
°C, and a solution of LiOH (6.9 kg, 164 mol) in water (157 L)
was added over 80 min at ∼15 °C. The resulting mixture was
agitated at ∼18 °C for three days17 and then heated to 82 °C
for 4.5 h and concentrated to remove 125 L of solvents at 23−
35 °C/10−60 mmHg. The concentrate was diluted with MTBE
(80 kg) and cooled to <8 °C. The pH was then cautiously
adjusted to 3.0 with conc. HCl (16.0 kg) at ∼8 °C. The organic
phase was separated and successively washed with water (61 L),
a mixture of water (61 L), and 15 wt % aq Na2CO3 (5.3 kg),
and a 2 wt % aq NaCl (60 kg). Then it was transferred to a
drum with a MTBE rinse (7.5 kg). The resulting solution of 47
in MTBE (∼83.3 kg, 53.4 mol in theory) was charged to a 200
L reactor and then concentrated by vacuum distillation at 15−
17 °C/130−165 mmHg until 70 L of solvent had been
removed. The concentrate was diluted with MTBE (40 kg),
and the distillation was resumed at 23 °C/60 mmHg until no
distillate was collected (62.5 L of solvent was removed) to give
47 (16.5 kg, 53.4 mol in theory) as an oil. Karl Fischer analysis
indicated the presence of 0.82 wt % water. This material was
used directly in the next step without further purification.
tert-Butyl-(4-acetoxy-6-fluoro-3-methylnaphthalen-2-
yl)acetate (49). The 200 L reactor, containing 47 (16.5 kg,
53.4 mol in theory) from above, was charged with Ac2O (42.3
kg), n-heptane (27 kg), and KOAc (7.74 kg, 79.0 mol). The
mixture was heated to 50 °C for 1.1 h and at 85−86 °C for 17
h; HPLC analysis indicated the presence of 0.12 area % of 48.
While the mixture was maintained at 19−25 °C, water (3 L)
was added in three portions over a period of 1.5 h. Then,
additional water (73 L) was added over 70 min. The resulting
suspension was agitated at 22−23 °C for 4.3 h and then
dropped into a centrifuge. The collected solids were washed
with potable water (5 L) and n-heptane (28 kg) and then dried
to give 49 (11.0 kg, 62% overall yield from 29, 98.8% HPLC
tert-Butyl-(6-fluoro-3-methyl-4-trifluoromethanesul-
fonyloxynaphthalen-2-yl)acetate (53). A 22 L flask was
charged with 50 (1.23 kg, 4.24 mol), CH2Cl2 (7.1 L), and
pyridine (686 mL, 8.48 mol). After the solution was cooled to 2
°C, Tf2O (857 mL, 5.09 mol) was added over 40 min at 7
5
°C. The resulting red solution was warmed to 18 °C, and
quenched with 1 M citric acid (5.1 L, 5.10 mol). The organic
layer was separated, washed with water (2 × 6.5 L), and
concentrated at 25 °C/100 mmHg to remove most of the
solvent. The residue was diluted with DMF (4.9 L) and further
concentrated. The remaining solution was warmed to 33 °C,
and water (4.9 L) was added over 1 h. The resulting suspension
was stirred at room temperature for 2.5 h and then filtered. The
filter cake was washed with DMF−water (1:1, v/v, 2 × 1.9 L)
and water (2 × 2.8 L) and dried to give 53 (1.76 kg, 98% yield)
1
as an off-white solid. Mp 87−88 °C; H NMR (400 MHz,
CDCl3) 7.81 (dd, J = 8.8, 5.3 Hz, 1H), 7.72 (s, 1H), 7.64 (dd, J
= 10.4, 2.6 Hz, 1H), 7.29 (td, J = 8.8, 2.6 Hz, 1H), 3.75 (s, 2H),
2.50 (s, 3H), 1.45 (s, 9H); 13C NMR (100 MHz, CDCl3)16
169.87, 162.82, 160.35, 142.50, 142.45, 132.47, 132.44, 130.24,
130.15, 130.10, 129.78, 129.27, 129.25, 127.27, 127.17, 123.43,
120.25, 117.26, 117.07, 117.00, 113.89, 105.40, 105.17, 81.85,
40.95, 27.91, 14.13; HRMS (ESI) m/z calcd for
C18H18O5F4NaS [M + Na]+ 445.0709, found 445.0704.
tert-Butyl-[6-fluoro-4-(4-methanesulfonylbenzyl)-3-
methylnaphthalen-2-yl]acetate (54). A 12 L flask was
charged with zinc dust (237 g, 3.63 mol), DMF (750 mL), and
iPrOH (8 mL, 0.10 mol). TMSCl (15.2 mL, 0.12 mol) was
added, and the suspension was stirred at room temperature for
30 min. Then, a solution of 11 (173 g, 0.84 mol) in DMF (375
mL) was added. The mixture was warmed to 30 °C to initiate
the reaction. Once the internal temperature reached 45 °C, the
remaining solution of 11 (173 g, 0.84 mol) in DMF (375 mL)
was added dropwise over 20 min at ∼45 °C. PdCl2(PPh3)2
(4.24 g, 6.04 mmol) was added, followed by 53 (500 g, 1.18
mol), and the mixture was warmed to 60 °C to initiate the
reaction. Without further heating, the internal temperature
increased to 86−95 °C over 10−15 min and then started to
decrease. After stirring at 65−85 °C for 2.5 h, the reaction
mixture was quenched with water (75 mL).
1
purity) as an off-white solid. Mp 150−151 °C; H NMR (300
MHz, CDCl3) 7.78 (dd, J = 9.1, 5.6 Hz, 1H), 7.61 (s, 1H), 7.31
(dd, J = 10.1, 2.6 Hz, 1H), 7.21 (td, J = 8.6, 2.6 Hz, 1H), 3.73
(s, 2H), 2.50 (s, 3H), 2.26 (s, 3H), 1.45 (s, 9H); 13C NMR
(100 MHz, DMSO-d6)16 170.09, 169.00, 161.77, 159.34,
143.75, 143.69, 132.62, 132.59, 130.58, 130.49, 129.07,
127.95, 126.69, 126.60, 116.23, 115.97, 104.91, 104.69, 80.42,
27.63, 20.41, 12.68; HRMS (ESI) m/z calcd for C19H21FO4Na
[M + Na]+ 355.1322, found 355.1311.
This reaction was repeated five times as described above, and
the reaction mixtures from all six runs were combined, diluted
with iPrOAc (15 L), and filtered through a pad of Celite (2.5
kg). The pad was washed with iPrOAc (3 × 5 L), and the
combined filtrates were transferred to an extractor. N-Acetyl-L-
cysteine (54.0 g, 0.33 mol) was added, and the mixture was
stirred at room temperature for 2 h, and then diluted with water
(7 L) and iPrOAc (2 L). The organic layer was separated and
washed successively with a mixture of DMF (8 L), 5 wt %
aqueous NaCl (8 L), and water (16 L), and then diluted with n-
heptane (15 L). The resulting suspension was filtered, and the
filter cake was washed with n-heptane (2 × 4 L) and dried to
give 54 (2.10 kg, 67% yield, 99.3% HPLC purity) as a white
solid. Additional 54 (865 g, 28% yield, 99.3% HPLC purity)
with identical purity was recovered from the filtrate after
additional aqueous washes followed by crystallization. Mp
tert-Butyl-(6-fluoro-4-hydroxy-3-methylnaphthalen-
2-yl)acetate (50). To a mixture of 49 (1.45 kg, 4.36 mol) and
methanol (8.7 L) was added potassium hydroxide (303 g, 5.40
mol) portionwise over 10 min. The reaction mixture was stirred
at 20−28 °C for 30 min, and then water (5.8 L) was added,
followed by the dropwise addition of conc. HCl (405 mL, 4.93
mol). The resulting suspension was filtered, and the filter cake
was washed successively with MeOH−water (1:1, 2 × 3.6 L)
and water (4.2 L), and dried to give 50 (1.23 kg, 97% yield) as
1
a white solid. Mp 131−132 °C; H NMR (300 MHz, DMSO-
1
d6) 9.09 (s, 1H), 7.86−7.76 (m, 2H), 7.35−7.26 (m, 2H), 3.71
(s, 2H), 2.24 (s, 3H), 1.41 (s, 9H); 13C NMR (100 MHz,
DMSO-d6)16 170.47, 160.74, 158.33, 149.21, 149.16, 132.87,
132.85, 130.05, 129.96, 129.21, 125.10, 125.01, 120.66, 120.64,
119.90, 115.36, 115.10, 105.35, 105.13, 80.15, 40.67, 27.68,
12.40; HRMS (ESI) m/z calcd for C17H18FO3Na [M + Na −
H]+ 312.1138, found 312.1128.
148−149 °C; H NMR (400 MHz, CDCl3) 7.82−7.78 (m,
3H), 7.68 (s, 1H), 7.38 (dd, J = 11.8, 2.6 Hz, 1H), 7.24 (m,
2H), 7.20 (td, J = 8.8, 2.6 Hz, 1H), 4.54 (s, 2H), 3.78 (s, 2H),
3.03 (s, 3H), 2.36 (s, 3H), 1.46 (s, 9H); 13C NMR (100 MHz,
CDCl3)16 170.83, 162.36, 159.93, 146.41, 138.40, 135.52,
132.95, 132.86, 131.94, 131.91, 131.66, 131.61, 130.83, 130.74,
129.32, 129.00, 128.88, 127.73, 115.57, 115.32, 107.59, 107.38,
656
dx.doi.org/10.1021/op300306c | Org. Process Res. Dev. 2013, 17, 651−657

Organic Process Research & Development
Article
(9) Fillion, E.; Fishlock, D.; Wilsily, A.; Goll, J. M. J. Org. Chem. 2005,
70, 1316.
(10) The lower purity and yield of 49 was possibly caused by the
relatively lower quality of sodium ethoxide compared to that of sodium
tert-butoxide.
81.20, 44.56, 41.83, 34.70, 28.04, 16.52; HRMS (ESI) m/z
calcd for C25H27O4FNaS [M + Na]+ 465.1512, found 465.1508.
[6-Fluoro-4-(4-methanesulfonylbenzyl)-3-methyl-
naphthalen-2-yl]acetic Acid (14). Three 22 L flasks were
each charged with 54 (989 g, 2.23 mol), water (495 mL),
AcOH (6.9 L), and conc. HCl (989 mL, 12.0 mol). Each
mixture was stirred at ∼47 °C for 90 min, and then water (8 L)
was added. The contents of the three flasks were poured into a
table-top funnel, and the collected solid cake was washed with
water (2 × 6 L) and dried to give 14 (2.47 kg, 95% yield, 99.5%
(11) Manolikakes, G.; Hernandez, C. M.; Schade, M. A.; Metzger, A.;
Knochel, P. J. Org. Chem. 2008, 73, 8422.
(12) (a) Erdik, E. Tetrahedron 1987, 43, 2203. (b) Deboves, H. J. C.;
Montalbetti, C. A. G. N.; Jackson, R. F. W. J. Chem. Soc., Perkin Trans.
1 2001, 1876.
(13) With the reversed addition order, the reaction worked
consistently on small scale, but it has not been tested on a large-
scale run.
(14) Proctor, L.; Dunn, P. J.; Hawkins, J. M.; Wells, A. S.; Williams,
M. T. Continuous Processing in the Pharmaceutical Industry. In Green
Chemistry in the Pharmaceutical Industry, Dunn, P. J., Wells, A. S.,
Williams, M. T., Eds; Wiley-VCH: Weinheim, Germany, 2010, pp
221−242.
(15) Compound 42 was agitated for 12 h to break up large, solid
chunks. This operation is not needed if fine powder of 42 is used.
(16) 13C NMR showed more peaks than the number of inequivalent
carbons due to F−C coupling.
(17) The hydrolysis of 45 was complete in 24 h. The reaction
mixture was held for an additional 2 days to avoid processing during
the weekend.
1
HPLC purity) as a white solid. Mp 209−210 °C; H NMR
(400 MHz, DMSO-d6) 12.46 (s, 1H), 7.94 (dd, J = 9.0, 6.2 Hz,
1H), 7.80 (d, J = 8.4 Hz, 2H), 7.79 (s, 1H), 7.68 (dd, J = 12.1,
2.3 Hz, 1H), 7.35 (td, J = 8.6, 2.3 Hz, 1H), 7.30 (d, J = 8.4,
2H), 4.60 (s, 2H), 3.84 (s, 2H), 3.16 (s, 3H), 2.33 (s, 3H); 13
C
NMR (100 MHz, DMSO-d6)16 172.71, 161.63, 159.22, 146.41,
138.54, 135.85, 132.66, 132.63, 132.41, 132.33, 132.28, 132.22,
130.98, 130.89, 129.01, 128.65, 127.22, 115.25, 115.00, 107.76,
107.55, 43.56, 39.99, 33.67, 16.23; HRMS (ESI) m/z calcd for
C21H19FO4NaS [M + Na]+ 409.0880, found 409.0880.
ASSOCIATED CONTENT
■
S
* Supporting Information
Copies of NMR spectra of compounds 14, 40, 43, 45, 47, 49,
50, 53, and 54. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: lshu@verizon.net.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are grateful to Dr. Masami Okabe for helpful discussion,
Ms. Rui Yu and Bojana Nedjic for analytical support, and Dr.
Hanspeter Michel for HRMS analysis. We thank Mr. Joseph
Eskow and his group for running steps 1−4 in the pilot plant.
REFERENCES
■
(1) Nagata, K.; Hirai, H.; Tanaka, K.; Ogawa, K.; Aso, T.; Sugamura,
K.; Nakamura, M.; Takano, S. FEBS Lett. 1999, 459, 195.
(2) For review on the development of CRTH2 antagonists, see
Pettipher, R.; Whittaker, M. J. Med. Chem. 2012, 55, 2915 and
references cited therein.
(3) Firooznia, F.; Gillespie, P.; Lin, T. A.; Mertz, E.; Sidduri, A.; So, S.
S.; Tan, J.; Thakkar, K. C. Naphthylacetic Acids. PCT Int. Pat. Appl.
WO/2010/055005, May 20, 2010; Chem. Abstr. 2010, 152, 591729.
Preliminary result of this investigation has been disclosed in this patent
application.
(4) Saito, S.; Shimada, I.; Takamori, Y.; Tanaka, M.; Maruoka, K.;
Yamamoto, H. Bull. Chem. Soc. Jpn. 1997, 70, 1671.
(5) Cruces, J.; Estev
Lett. 2001, 42, 4825.
́ ́
ez, J. C.; Castedo, L.; Estevez, R. J. Tetrahedron
(6) (a) Caron, S.; Do, N. M.; Sieser, J. E.; Arpin, P.; Vazquez, E. Org.
Process Res. Dev. 2007, 11, 1015. (b) Maillard, J.; Langlois, M.;
Delaunay, P.; Vo, V. T.; Meingan, J. P.; Rapin, M.; Morin, R.; Manuel,
C.; Mazmanian, C. Eur. J. Med. Chem. 1977, 12, 161.
(7) (a) Davies, K. A.; Abel, R. C.; Wulff, J. E. J. Org. Chem. 2009, 74,
3997. (b) Bieber, L. W.; da Silva, M. F. Tetrahedron Lett. 2007, 48,
7088. (c) Hooz, J.; Calzada, J. G.; McMaster, D. Tetrahedron Lett.
1985, 26, 271.
(8) Lang, R. W.; Hansen, H.-J. Org. Synth. 1984, 62, 202.
657
dx.doi.org/10.1021/op300306c | Org. Process Res. Dev. 2013, 17, 651−657