Development of a potential manufacturing route to PF-00610355: A novel inhaled β2-adrenoreceptor agonist

Article abstract of DOI:10.1021/op2002408

The development of a practical, scalable route to PF-00610355 (8) is described. In this convergent approach, amine 9 is coupled to protected bromohydrin 1 to give the doubly protected intermediate 26. TBS-Deprotection of 26 affords the benzyl protected pe

Full text of DOI:10.1021/op2002408

ARTICLE
pubs.acs.org/OPRD
Development of a Potential Manufacturing Route to PF-00610355: A
Novel Inhaled β2-Adrenoreceptor Agonist
Pieter D. de Koning,* Nieves Castro, Iain R. Gladwell, Natalie A. Morrison, Ian B. Moses,
Maninder S. Panesar, Alan J. Pettman, and Nicholas M. Thomson
Pfizer Global Research and Development, Sandwich Laboratories, Ramsgate Road, Sandwich, Kent CT13 9NJ, United Kingdom
ABSTRACT: The development of a practical, scalable route to PF-00610355 (8) is described. In this convergent approach, amine 9
is coupled to protected bromohydrin 1 to give the doubly protected intermediate 26. TBS-Deprotection of 26 affords the benzyl
protected penultimate intermediate 25 which is crystallized as the corresponding hemifumarate salt 25a. On the basis of solubility
data, the final debenzylation was conducted in aqueous THF, and the API (8) is isolated from acetonitrile by an unusual distillative
crystallization process. The development of an efficient process to prepare amine 9 is also described.
’ INTRODUCTION
Both the medicinal chemistry1 and our initial scale-up2 routes
had used a broadly similar strategy, wherein first the headgroup 1
and linker 2 were combined, and the tail 7 was added last. From a
simple step-count analysis, this was clearly suboptimal, as 7 is
prepared in 3 steps while 1 requires 6 steps. Additionally, from an
evaluation of the literature, the preferred routes to prepare other
β2-agonists are those wherein the key CꢀN bond is formed using
a fully elaborated ‘linker-tail’ portion (e.g., amine 9), as depicted
in the retrosynthesis of 8 in Scheme 2. In particular, salmeterol is
prepared from a protected chiral ethanolamine fragment and an
electrophile5 (Scheme 2, route a) and both formoterol6 and
indacaterol7 from the reaction between an epoxide and a fully
elaborated amine (Scheme 2, route b). In the case of 8, the
former route was considered less likely to succeed due to
the challenge of introducing the quaternary carbon adjacent to
the secondary amine,8 and priority was given to exploration of
the latter strategy, focusing on the well-precedented epoxide or
halohydrin routes shown in Scheme 2.9
The use of long-acting inhaled β2-adrenoreceptor agonists is
an established therapy for the treatment of respiratory diseases
such as asthma and chronic obstructive pulmonary disease
(COPD). As part of a program to identify an ultralong acting
β2-adrenoreceptor agonist suitable for once-daily dosing, N-[(40-
hydroxybiphenyl-3-yl)methyl]-2-[3-(2-{[(2R)-2-hydroxy-2-{4-
hydroxy-3-[(methylsulfonyl)amino]phenyl}ethyl]amino}-2-
methylpropyl)phenyl]acetamide (PF-00610355, 8), was progressed
into clinical development.1 We have recently disclosed the initial
synthetic route used to prepare material for early development
studies (Scheme 1).2 While this route was ultimately successful,
the final step and subsequent purification of 8 proved extremely
challenging and despite considerable efforts, only provided a
modest 36% yield of clinical quality material. Consequently, we
initiated work on the identification of an alternative route that
would be more suitable for large-scale manufacture.
Synthesis of Amine 9. In order to examine these alternative
approaches, supplies of amine 9 were required. Since amino ester
2 was available from the previous campaign (as the di-(p-toluoyl)-
(L)-tartaric acid (DTTA) salt, prepared in 7 steps3), this was
converted to amine 9 as shown in Scheme 3. Boc-protection and
ester hydrolysis gave protected amino acid 10 which was coupled
’ RESULTS AND DISCUSSION
The initial medicinal chemistry design of 8 was based on a
modular construct, comprising three components, a sulfonamide
‘headgroup’ 1, an amino ester ‘linker’ 2, and an amine ‘tail’ 7
(Scheme 2). Given the demand for material and the limited time
available, as efficient processes to all three components had already
been developed,2,3 it was evident that any new process would need
to utilize these or very closely related building blocks. Consequently,
the key considerations during the initial analysis phase were to
define the optimal route by which to combine the fragments, choice
of protecting groups,4 and critically, the final step. On the basis of
our hard earned knowledge of the challenges of purifying 8,2 it was
clear that the final step would need to be mild, high-yielding, and
chemoselective. In addition, the penultimate intermediate would
need to be readily prepared in high purity. This led us to two
possible options, either a final step debenzylation or a final step
desilylation. Both of these are highly chemoselective processes that
occur under mild conditions and were unlikely to introduce
additional structurally related impurities, although each presents
significant processing challenges as discussed herein.
with amine 7 HCl using EDC in the presence of diisopropylethy-
3
lamineand catalyticDMAP. Finally, Boc deprotection affordedthe
desired amine 9. While this route successfully provided the initial
supplies of 9 it was rather lengthy (10 steps) as a result of multiple
protecting group interchanges, and ultimately a more efficient
synthesis was required.
Since we had already developed an efficient route to chlor-
oacetamide 13,3 initial studies focused on coupling 13 with amine 7.
Despite considerable effort, this was unsuccessful, largely due to
the reactivity of the chloroacetamide group, resulting in the
formation of numerous, unidentified byproducts under a range of
amide bond-forming conditions (e.g., CDI, EDC, DCC, isobutyl
Received: August 31, 2011
Published: September 30, 2011
r
2011 American Chemical Society
1256
dx.doi.org/10.1021/op2002408 Org. Process Res. Dev. 2011, 15, 1256–1265
|

Organic Process Research & Development
ARTICLE
Scheme 1. Initial synthetic route to 8a
a Reagents and conditions: (a) propionitrile (EtCN), reflux; (b) Et3N 3HF, EtCN; (c) (i) NaOH, EtCN/water, then HCl/1,4-dioxane; (ii) water
3
reslurry, 80%; (d) NaOH, H2, 20% Pd(OH)2/C, water, then acetonitrile, 69%; (e) (i) EDC HCl, pyridine, then water, 80%; (ii) acetone/water; (iii)
MeOH/water reslurry, 45%.
3
Scheme 2. Summarized retrosynthetic analysis
chloroformate). Having established that the chloroacetamide
was not a suitable protecting group, we identified a literature
precedent for preparation and cleavage of a trichloroacetamide
via a Ritter reaction of an alcohol with trichloroacetonitrile,
followed by basic hydrolysis.10
The Ritter reaction between alcohol 12 (prepared in 4 steps3)
and trichloroacetonitrile proceeded reasonably well, affording
the desired trichloroacetamide 14 (83% crude yield). Subsequent
coupling to amine 7 gave the desired amide 15 and basic hydrolysis
afforded 9 in modest overall yield (27%, Scheme 4).11 Upon
1257
dx.doi.org/10.1021/op2002408 |Org. Process Res. Dev. 2011, 15, 1256–1265

Organic Process Research & Development
ARTICLE
Scheme 3a Initial route to amine 9
Scheme 5a Optimized route to amine 9
a Reagents and conditions: (a) Et3N, EDC HCl, HOBt, CH2Cl2; (b)
3
chloroacetonitrile, TFA, 50 °C; (c) thiourea, AcOH, reflux, then
acetonitrile cryst, 64%.
a Reagents and conditions: (a) (i) Boc2O, i-Pr2NEt, EtCN; (ii) NaOH,
Initial studies on the Ritter reaction of alcohol 16 with
chloroacetonitrile using the conditions developed previously
for alcohol 12 (H2SO4 and AcOH in dichloromethane)3 were
promising; however, as the reaction scale increased we observed
increasing amounts of the intermediate olefin arising from
dehydration of alcohol 16, and were unable to convert this to
the desired 17. A literature survey identified trifluoroacetic acid
(TFA)12 as an alternative to the commonly used sulfuric acid13 in
the Ritter reaction. To our delight, when we examined the Ritter
reaction of 16 with chloroacetonitrile in TFA, full conversion to
the chloroacetamide 17 was observed after 2 h at 50 °C.14
Chemoselective deprotection of chloroacetamide 17 was
accomplished by treatment with thiourea in acetic acid,15 afford-
ing pure amine 9 after removal of the byproduct 18 and crystal-
lization (Scheme 5). As the target amine 9 was known to be a
crystalline solid and it proved challenging to crystallize the
intermediate alcohol 16 and chloroacetamide 17, we decided
to develop a fully telescoped process from alcohol 12 to amine 9.
The preferred solvent for the initial amide formation was
dichloromethane, however, during the workup solubility pro-
blems were encountered (an oily phase separated out of solution
during the aqueous washes); this was overcome by dilution with
THF once the reaction was complete. After an aqueous workup,
the solvent was readily exchanged to chloroacetonitrile (bp
124ꢀ126 °C) by distillation. Treatment with TFA at 50 °C
afforded the desired chloroacetamide 17. Unfortunately, thiour-
ea deprotection proved unsuccessful in this TFA/chloroacetoni-
trile mixture; consequently dilution with dichloromethane and
an aqueous workup was required to remove the TFA. Thereafter,
the dichloromethane was removed by distillation and replaced
with acetic acid; subsequent treatment with thiourea at reflux
affected the desired deprotection. After filtration to remove
byproduct 18 (Scheme 5), the acetic acid solution of amine 9
was diluted with water and back-extracted to remove residual
chloroacetonitrile.16 The workup was completed by neutralization
and extraction into 2-methyltetrahydrofuran, followed by solvent
exchange to acetonitrile from which 9 was isolated as a crystalline
solid in an acceptable 64% yield from 12 on pilot plant scale.
Conversion of amine 9 to API (8). In parallel with the
development of the route to amine 9, both the epoxide and
halohydrin routes to the active pharmaceutical ingredient (API)
8 were evaluated. While the halohydrin route offered the option
of either a final step debenzylation or desilylation, the epoxide
route only offered the former; however, the epoxide approach
THF/water, then HCl; (iii) toluene/heptane cryst, 90%; (b) (i) 7 HCl,
3
i-Pr2NEt, DMAP, EDC HCl, MeCN; (ii) water; (iii) aq citric acid, 82%;
3
(c) (i) TFA/CH2Cl2, then water/aqueous ammonia; (ii) acetone
cryst, 54%.
Scheme 4aAlternative routes to amine 9
a Reagents and conditions: (a) chloroacetonitrile, H2SO4, AcOH,
CH2Cl2; 65ꢀ70%; (b) trichloroacetonitrile, H2SO4, AcOH, then hep-
tane cryst, 83%; (c) 7 HCl, Et3N, EDC HCl, HOBt, EtOAc; (d) KOH,
3
3
EtOH/water, then acetone cryst, 32%.
scale-up to around 100 g of alcohol 12, the overall yield
plummeted to <10%. While additional development work might
have succeeded in identifying the problems and improving the
yield, we had some concerns about generating chloroform in the
deprotection step, and this, coupled with the successful devel-
opment of an alternative route (see Scheme 5) meant that this
process was not investigated further.
Recognizing that deferring installation of the amine function-
ality until after the amide bond formation step would avoid
the issues encountered with the reactive chloroacetamide, we
decided to examine the step-reordered sequence shown in
Scheme 5. A range of conditions (e.g., CDI, DCC, EDC, and
isobutyl chloroformate) for the coupling of amino alcohol 12
with amine 7 were examined. From this screen, EDC HCl in the
3
presence of HOBt and Et3N emerged as the preferred reaction
conditions.
1258
dx.doi.org/10.1021/op2002408 |Org. Process Res. Dev. 2011, 15, 1256–1265

Organic Process Research & Development
ARTICLE
does not require the TBS-protecting group, is potentially shorter,
and has been successfully used for the synthesis of both for-
moterol and indacaterol.6,7
bromohydrin 1 will not generate any regioisomeric impurity
(formed due to epoxide opening at the benzylic position), and
the presence of the TBS group as well as the bulky amine 9
should disfavor overalkylation and the formation of ‘dimeric’
species. The disadvantages of this approach are the slightly longer
synthesis of 1,19 a slower coupling reaction, and the cost of the
TBS protecting group.
Bromohydrin 19 was synthesized as described,2 and treatment
with potassium carbonate in THF/MeOH afforded the antici-
pated epoxide 20; however, this rapidly degraded during the
workup and isolation process (Scheme 6). Attempts at generating
and reacting epoxide 20 in situ were also unsuccessful. Postulating
that the instability might be related to the acidic sulfonamide
proton, this was protected by addition of a second methanesulfo-
namide group. For these initial proof-of-concept studies, epoxide
23waspreparedfromTBSbromohydrin1,2 asshowninScheme6.
Addition of the second methanesulfonamide group and deprotec-
tion afforded bromohydrin 22, and treatment with base smoothly
converted 22 to the desired epoxide 23, which was readily isolated
and characterized.17
As previously noted when conducting a similar coupling reaction,2
some TBS-deprotection of the product 26 occurred under the
reaction conditions. On the basis of our previous work,2 this coupling
reaction was initially conducted in EtCN with 2 equiv of 9; however,
switching to the higher-boiling butyronitrile (bp 118 °C) resulted in a
significant increase in reaction rate (consumption of starting material
within 24ꢀ36 h on small scale, rather than >48 h). Recognizing that
the use of 2 equiv of 9was inefficient and was likely to be prohibitively
expensive in the long term, an extensive screen of alternative bases was
conducted.20 From this screen, addition of 5 equiv of sodium
hydrogen carbonate (NaHCO3) was found to completely suppress
TBS-deprotection of 26, but the reaction mixture proved challenging
to stir. The best compromise was found to be 3 equiv of NaHCO3
and 1 equiv of 9 in refluxing butyronitrile; under these conditions the
coupling proceeded to completion in around 30 h with minimal TBS
deprotection on lab scale. The bis-protected product 26 was not
crystalline (and a crystalline form was not identified from an extensive
salt screen) and was initially purified by chromatography, but
subsequently the crude solution of 26 was simply telescoped into
the next step.
Reaction of epoxide 23 with amine 9 proceeded smoothly,
although elevated temperatures were required to reach comple-
tion within 24 h (Scheme 7). After some screening, the best
solvent was identified as butyronitrile (bp 118 °C); however,
even under these conditions, significant byproduct formation
was noted.18 The product 24 was isolated by chromatography in
54% yield, and treatment with aqueous sodium hydroxide
afforded 25 in a moderate 60% yield.
In contrast, although sluggish, the reaction between TBS
bromohydrin 1 and amine 9 in refluxing propionitrile (EtCN,
bp 97 °C)2 proceeded to completion in around 54ꢀ60 h, with
high selectivity and minimal byproduct generation. The reduc-
tion in byproduct levels was anticipated since, unlike epoxide 23,
At this stage it was apparent that, while the epoxide route was
potentially slightly shorter, the need for additional protection of
the sulfonamide coupled with the relatively poor reaction profile,
lower yields, and reduction in endgame flexibility meant that the
halohydrin route was our preferred option. Having decided to
utilize the halohydrin route, the use of butyronitrile for the
coupling step was re-evaluated since it is not currently ICH-
listed21 and using it at the end of the synthesis would require
justification. A range of alternative, ICH-listed solvents was
screened, and n-butyl acetate (bp 126 °C) was identified as a
suitable replacement. Due to the poor solubility of amine 9 in n-
butyl acetate at ambient temperature, concentrated reaction
mixtures (around 3 mL/g with respect to amine 9) proved
impossible to stir (at 20ꢀ25 °C); however, using higher dilution
resulted in extremely slow reaction. This was overcome by
preparing the initial reaction mixture in a relatively high volume
of n-butyl acetate (10 mL/g with respect to amine 9), affording a
mobile slurry. Upon heating to reflux, the solubility of 9 was
greatly increased, and concentration of the reaction mixture by
distillation of around 7ꢀ8 mL/g of n-butyl acetate resulted in a
significant rate increase while still keeping the reactants (1 and 9)
in solution. A slight excess of either amine 9 or TBS bromohydrin
1 could be tolerated, but since bromohydrin 1 was easier to purge,
Scheme 6a Preparation of epoxides
a Reagents and conditions: (a) MsCl, i-Pr2NEt, MeCN, then water;
(b) Et3N 3HF, THF/MeOH, then EtOAc, 73%; (c) K2CO3, THF/
3
MeOH; dec (for 20) or 100% (for 23).
Scheme 7a Conversion of epoxide 23 to 25
a Reagents and conditions: (a) 9, butyronitrile, reflux, then chromatography, 54%; (b) NaOH, EtOH/water, 60%.
1259
dx.doi.org/10.1021/op2002408 |Org. Process Res. Dev. 2011, 15, 1256–1265

Organic Process Research & Development
ARTICLE
Scheme 8. Coupling of amine 9 and bromide 1 and conversion to 8a
a Reagents and conditions: (a) 9, n-butyl acetate, NaHCO3, reflux; (b) H2, Pd(OH)2/C, n-butyl acetate/ethyl acetate; (c) Et3N 3HF, n-butyl acetate/
3
ethyl acetate/MeOH; (d) NH4F, MeOH/water; (e) H2, Pd/C, THF/water.
an excess (1.05 equiv) of this was used. Sodium hydrogen
carbonate (3 equiv) remained the ideal HBr scavenger under
these conditions.22 On pilot-plant scale the reaction took around
40 h to reach adequate conversion (defined as <5% 9 remaining).
Once the reaction was complete, the mixture was cooled to 50 °C
and diluted with ethyl acetate,23 prior to cooling to 20 °C. Initially,
a series of acidic, basic, and water washes was employed to remove
residual amine 9 and inorganic salts; this was simplified to a single
water wash to remove the inorganic salts once adequate purge of 9
was demonstrated in the downstream steps.
In order to decide between the two potential endgame
sequences (Scheme 8), two major criteria needed to be assessed.
First, the final step needed to be thoroughly evaluated to ensure
that suitable quality 8 could be isolated. Second, the feasibility of
isolation and purification of the immediate precursors, benzyl-
protected 25 or TBS-protected 27, preferably by crystallization,
needed to be established.
In order to evaluate the final step desilylation option
(Scheme 8, steps b and d), the fully protected 26 was readily
debenzylated by hydrogenolysis in the presence of palladium on
carbon. Conveniently, the crude solution of 26 after workup (in a
mixture of ethyl acetate and n-butyl acetate) could be used for
this purpose. The product, 27, was isolated as a foam after
chromatographic purification. A large salt screen was conducted
on 27 in an attempt to identify a crystalline form, however this
was unsuccessful. Our medicinal chemistry colleagues had de-
protected 27 using a large excess of ammonium fluoride (10
equiv) in aqueous ethanol,1 and we were particularly attracted to
the fact that this was a direct-drop process24 that afforded
acceptable quality 8 directly from the reaction mixture. Since
we were concerned about the use of ammonium fluoride, which is
both toxic and incompatible with glass, a selection of alternative
deprotection conditions were examined (e.g., TBAF, triethyla-
mine trihydrofluoride). While all of these reagents smoothly
deprotected 27, the product (8) remained in solution25 and
removing the residual reagent and byproducts proved challenging
due to the low solubility (in common extraction solvents) and
instability of 8. Attempts at reducing the number of equivalents of
NH4F were also unsuccessful, as the precipitation appears to
require a high ionic strength to work effectively.
Since both the coupling and deprotection steps proceeded
with minimal byproduct generation, we examined the possibility
of telescoping the crude debenzylation product (27) into the
desilylation step. This proved successful, affording pure 8 in an
acceptable 50% yield from amine 9 (Scheme 8).
To evaluate the final-step debenzylation option (Scheme 8,
steps c and e), 26 was deprotected by treatment with Et3N 3HF,
3
affording 25 as a foam. Initial salt screening experiments identi-
fied a selection of amorphous salts of 25 that could be precipitated
by addition of a solution of the salt in ethanol or a mixture of
2-butanone (methyl ethyl ketone, MEK) and n-butyl acetate to
MTBE; the best of these was an amorphous dibenzoyl-(L)-tartaric
acid salt. Given the potential challenges of developing a robust
process to an amorphous salt, when further screening work led to
the identification of a crystalline hemifumarate salt 25a (Scheme 9),
this was selected for development.
The final debenzylation reaction was complicated by the low
solubility of 8 in most common hydrogenation solvents (e.g.,
MeOH), with the product precipitating from solution during the
reaction, making separation from the heterogeneous hydrogena-
tion catalyst challenging. While the reaction proceeded smoothly
in solvents like DMF and NMP (in which 8 is sufficiently soluble),
significant levels of Pd leaching were observed. Additionally, while
8 could be precipitated from these solvents by addition of water,
the recovery was low, and the isolated 8 was not suitable for
downstream processing. Alternative workup procedures were also
unsuccessful due to the low solubility of 8 in water-immiscible
extraction solvents.
Extensive solubility screening of 8 identified that it was
reasonably soluble (∼10 mL/g) in a 9:1 mixture of THF and
water. Hydrogenolysis of 25 in this mixture afforded a solution of 8;
however, isolation of crystalline material with appropriate
1260
dx.doi.org/10.1021/op2002408 |Org. Process Res. Dev. 2011, 15, 1256–1265

Organic Process Research & Development
ARTICLE
Scheme 9a Optimized process to 8
a Reagents and conditions: (a) 9, n-butyl acetate, NaHCO3, reflux, then ethyl acetate/water; (b) Et3N 3HF, n-butyl acetate/ethyl acetate/MeOH;
3
(c) fumaric acid, n-butyl acetate/MEK, 87%; (d) (i) aqueous ammonia, THF/water; (ii) carbon filtration; (iii) H2, Pd/C, THF/water; (iv) azeotropic
distillation to MeCN; (v) MeOH/water, 50 °C, 70%.
solid-form properties for an inhaled therapeutic agent26 proved
challenging. Since 8 is insoluble in both water and anhydrous
THF, adjusting the solvent composition of the initial product
solution by distillation to pure water or THF was examined.
Unfortunately, this only afforded low yields of poor-quality
product that was unsuitable for downstream processing. Dilution
with excess water caused the product to oil out of solution, and
addition of other antisolvents (e.g., methanol, ethanol, heptane)
was equally unsuccessful.
the residue was dissolved in MEK. Addition of a small amount of
water was essential to ensure a homogeneous solution was
obtained. Upon addition of fumaric acid (0.5 equiv), an oily
residue separated out of solution, but upon extended reflux
(around 4 h), this slowly converted to the desired crystalline
hemifumarate salt 25a. Even in the presence of seed material, this
initial oily phase could not be avoided; however, the process to
convert it into a crystalline form proved robust and was success-
fully implemented on pilot-plant scale, affording >40 kg of 25a in
∼85% yield.
During the development of the previous process to prepare 8,2
wehad evaluatedanextensive range of solventsand solventꢀwater
combinations for the purification of 8. While it had not proved
useful for purification of 8, we had noted that acetonitrile was a
particularly good antisolvent; this, coupled with the knowledge
that azeotropic distillation with acetonitrile would be an effective
way to remove both THF and water, led us to attempt to isolate 8
from acetonitrile. As anticipated, azeotropic distillation with
acetonitrile was successful in removing both THF and water
and, to our delight, afforded a mobile slurry of 8. Upon isolation,
the product was found to have suitable physical properties for
downstream processing, but the purity was slightly lower than
required. Fortunately, a reslurry in aqueous methanol provided
sufficient purity upgrade without affecting the physical properties.
Having examined both endgame options, we selected the final-
step debenzylation route for further development as a crystalline
penultimate intermediate (25a) and viable conditions for the
final step had been identified. In contrast, while the final-step
desilylation route had been demonstrated to provide suitable
quality 8, purification of 27 required chromatography.
Since the hydrogenolysis had to be conducted in a 9:1 THF/
water mixture to ensure the product 8 was soluble, the fact that
25 was isolated as a salt posed a concern. As THF is miscible with
water, conducting a salt break step with an aqueous base could
result in THF with a variable water content and pose the risk of
inorganic contamination of the API. While it would have been
possible to first conduct the salt break in a different solvent (e.g.,
dichloromethane) and then exchange this to THF for the
hydrogenation step, we ideally wanted to use a single solvent.
An additional concern was the possibility of deprotonating the
sulfonamide; thus, only weak bases were evaluated.
After some experimentation, the best option was identified as
treating a THF suspension of 25a with aqueous ammonia of
sufficient ionic strength to be immiscible with THF. The result-
ing THF solution of 25 was then azeotropically distilled to
remove water and any residual ammonia, giving a THF solution
of 25. Subsequent dilution with water provided the appropriate
solvent composition for the hydrogenolysis step.
Initially, the hydrogenolysis proceeded to completion within a
few hours; however, once we started using fumarate 25a derived
from amine 9 that had been prepared using the optimized
process (Scheme 5), several batches failed and required an
additional charge of catalyst to reach full conversion. This
suggested that a catalyst poison had been introduced, and from
an inspection of the route, the most likely source was the thiourea
deprotection step (Scheme 5, conversion of chloroacetamide 17
to amine 9), with the catalyst poison tracking through the
subsequent steps. Fortunately, treatment with activated carbon27
removed the catalyst poison,28 and the hydrogenolysis pro-
ceeded to completion with a single catalyst charge, affording 8
in 70% yield on pilot-plant scale, after isolation and purification.
As this was the only crystalline intermediate in the sequence, a
telescoped process through to fumarate salt 25a was developed
(Scheme 9). The coupling reaction was conducted as described
previously, affording a solution of silyl ether 26 in a mixture of
n-butyl acetate and ethyl acetate. Treatment with Et3N 3HF
3
removed the TBS group; however, some material was observed
to oil out of solution as the reaction progressed. Fortunately,
addition of a small amount of methanol to the reaction mixture
solubilized this material, resolving the problem.
Once the reaction was complete, excess reagent was destroyed
by addition of aqueous ammonia. The solution was concentrated
to remove the volatile solvents (ethyl acetate and methanol), and
1261
dx.doi.org/10.1021/op2002408 |Org. Process Res. Dev. 2011, 15, 1256–1265

Organic Process Research & Development
ARTICLE
The use of palladium in the final step did pose a concern, and the
palladium level in the isolated 8 was monitored to ensure that it
was within safe limits. If the palladium level breached the 20
ppm limit, treatment of a THF/water solution of 8 with Quad-
rapure TU resin, followed by the usual isolation from acetonitrile
and methanol/water reslurry, successfully purged palladium
without significant degradation of 8.29
In conclusion, herein we describe the development of a
practical, scalable route to PF-00610355 (8), used to prepare
>10 kg of material with the potential to form the basis of a larger-
scale process to this complex molecule. In this convergent
approach, a fully elaborated amine 9 is coupled to protected
bromohydrin 1 to give the fully protected intermediate 26 which
is not isolated and is converted to benzyl-protected penultimate
intermediate 25 and crystallized as the hemifumarate salt 25a. On
the basis of solubility data, the final debenzylation is conducted in
aqueous THF, and the API (8) is isolated from acetonitrile by an
unusual distillative crystallization process. The development of
an efficient process to prepare amine 9 is also described.
filtration, washed with water (100 mL), and dried at 40 °C under
vacuum to give the title compound 11 (31 g; 82%) as a white solid.
1H NMR (DMSO-d6, 400 MHz) δ: 9.50 (1H, s), 8.54 (1H, t, J =
5.9 Hz), 7.44ꢀ7.39 (4H, m), 7.33 (1H, t, J = 7.6 Hz), 7.21ꢀ7.11
(3H, m), 7.06 (1H, s), 7.00 (1H, m), 6.84 (2H, d, J = 8.6 Hz), 6.25
(1H, s), 4.30 (2H, d, J = 6.1 Hz), 3.46 (2H, s), 2.88 (2H, s), 1.45
(9H, s), 1.14 (6H, s). 13C NMR (100 MHz, DMSO-d6) δ: 170.1,
157.1, 154.3, 140.2, 139.9, 138.3, 135.7, 131.2, 130.8, 128.7, 128.4,
127.6, 127.5, 126.6, 125.2, 124.7, 124.4, 115.7, 77.1, 52.1, 48.7, 42.4,
42.2, 28.4, 27.1, 26.8. LCMS: Found m/z 489.27 [M + H]+.
2-[3-(2-Amino-2-methylpropyl)phenyl]-N-[(40-hydroxybi-
phenyl-3-yl)methyl]acetamide 9. Method A. A suspension of
Boc amine 11 (31.0 g; 63.4 mmol) in dichloromethane (150 mL)
was stirred under an inert atmosphere while trifluoroacetic acid
(50 mL; 649 mmol) was added. The resulting pale orange-brown
solution was stirred for 1.5 h, then concentrated under reduced
pressure to give a thick brown oil. The oil was treated with a
mixture of water and concentrated aqueous ammonia (9:1,
∼250 mL) until pH 12 was reached, and then the mixture was
extracted with a mixture of ethyl acetate and methanol (9:1, 2 ꢁ
150 mL). The combined organic extracts were washed with water
and concentrated under reduced pressure. The resulting foam
was refluxed in acetone (500 mL) for 1 h, and the resulting slurry
was cooled to ambient temperature and stirred overnight. The
solid was isolated by filtration, washing with acetone, and dried at
40 °C in a vacuum oven to give the title compound 9 (13.4 g;
54%) as a white solid. Mp 123 °C; 1H NMR (400 MHz, DMSO-d6)
δ: 8.59 (1H, t, J = 5.9 Hz), 7.41 (4H, m), 7.32 (1H, t, J = 7.4 Hz),
7.24ꢀ7.12 (4H, m), 7.06 (1H, br d, J = 7.2 Hz), 6.86 (2H, dm, J =
8.6 Hz), 4.34 (2H, d, J = 5.9 Hz), 3.49 (2H, s), 2.55 (2H, s), 0.97
(6H, s). 13C NMR (100 MHz, DMSO-d6) δ: 170.3, 157.3, 140.2,
139.9, 138.7, 135.8, 131.0, 130.7, 128.7, 128.4, 127.6, 127.5, 126.6,
125.2, 124.7, 124.3, 115.8, 50.4, 49.8, 42.5, 42.2, 29.9. LCMS: Found
’ EXPERIMENTAL SECTION
tert-Butyl-{1-[3-(2-{[(40-hydroxybiphenyl-3-yl)methyl]amino}-
2-oxoethyl)phenyl]-2-methylpropan-2-yl}carbamate 11. Diiso-
propylethylamine (210 mL; 1.21 mol) was added to a suspension of
2 DTTA3 (250 g; 0.40 mol) in propionitrile (1.0 L), giving a pale-
3
yellow solution. A solution of di-tert-butyl dicarbonate (97 g; 0.44
mol) in propionitrile (250 mL) was added, and the resulting pale-
yellow solution was stirred at ambient temperature for 21 h. Water
(250 mL) was added, and the mixture was stirred for 30 min. The
phases were separated, and the organic phase was washed successively
with 10% aqueous citric acid (500 mL), water (300 mL), saturated
aqueous sodium hydrogen carbonate (500 mL), and brine (500 mL).
The organic phase was concentrated to a dark-orange oil and dissol-
ved in a mixture of tetrahydrofuran (250 mL) and water (250 mL).
Sodium hydroxide (80 g; 2.0 mol) was added, and the resulting
mixture was stirred at ambient temperature for 91 h. Toluene
(400 mL) was added, and the mixture was stirred for 30 min; then
the phases were separated. The organic phase was extracted with a
mixture of water (200 mL) and saturated aqueous sodium hydrogen
carbonate (100 mL). The combined aqueous phase was adjusted to
pH 1 with concentrated hydrochloric acid and extracted with ethyl
acetate (2 ꢁ 250 mL). The combined ethyl acetate extracts were
washed with water (2 ꢁ 200 mL) and then concentrated to dryness.
The resulting oil was dissolved in refluxing toluene (100 mL), and
heptane (∼400 mL) was added. The mixture was cooled to ambient
temperature and stirred for 3 h. The solid was isolated by filtration,
washing with heptane (2 ꢁ 200 mL) and dried in a vacuum oven at
40 °C to give Boc amino acid 10 (111 g; 90%) as a pale-yellow solid.
1H NMR (CDCl3, 400 MHz) δ: 7.23 (1H, m), 7.15 (1H, m), 7.07
(2H, m), 3.61 (2H, s), 2.96 (2H, s), 1.47 (9H, s), 1.25 (6H, s). A
mixture of Boc amino acid 10 (25 g; 81.3 mmol), amine hydro-
m/z 389.27 [M + H]+. Anal. Calcd For C25H28N2O2 1/2(H2O):
3
C, 75.54; H, 7.35; N, 7.05. Found: C, 75.48; H, 7.05; N, 7.00.
Method B. Triethylamine (22.6 L; 163.2 mol) was added to a
mixture of alcohol 12 (17 kg; 81.6 mol), amine hydrochloride
7 HCl (21.1 kg, 89.5 mol), and 1-hydroxybenzotriazole hydrate
3
(5.5 kg; 40.8 mol) in dichloromethane (155 L) at 20 °C under
nitrogen, followed by a dichloromethane line wash (15 L). The
mixture was stirred at 20 °C for 1 h, then 1-(dimethylaminopropyl)-
3-ethylcarbodiimide hydrochloride (15.6 kg, 81.6 mol) was added,
and the mixture was stirred for 5 h, at which point HPLC analysis
showed complete conversion to amide 16. THF (85 L) and water
(85 L) were added, and after 10 min the phases were separated. The
retained lower organic phase was washed with water (85 L), 1 M
aqueous HCl (2 ꢁ 85 L), and 1 M aqueous potassium bicarbonate
(2 ꢁ 85 L). The organic solution was diluted with chloroacetonitrile
(61.2 L) and concentrated by distillation until a vapor temperature
of 89 °C was reached (approximately 150 L of distillate was
collected). The solution was cooled to 50 °C, and trifluoroacetic
acid (153 L) was added over 2 h, maintaining the temperature at
50 °C. Once the addition was complete, the solution was stirred at
50 °C for a further 10 h, at which point HPLC analysis indicated
complete conversion to chloroacetamide 17. The solution was
cooled to 20 °C and diluted with dichloromethane (153 L). The
solution was washed with water (2 ꢁ 306 L) and 1 M aqueous
potassium bicarbonate (2 ꢁ 153 L)30 and then diluted with acetic
acid (187 L). The solution was concentrated by distillation until a
vapor temperature of 92 °C was reached (approximately 150 L of
distillate was collected), and then cooled to 20 °C. Thiourea (26.1 kg;
342.7 mol) was added, and the mixture was heated to 70 °C and held
chloride 7 HCl (18.2 g; 77.3 mmol), 4-dimethylaminopyridine
3
(100 mg; 0.81 mmol), and diisopropylethylamine (22.1 g; 170.8
mmol) in acetonitrile (125 mL) was stirred at ambient temperature
under nitrogen while 1-(3-dimethylaminopropyl)-3-ethylcarbodii-
mide hydrochloride (17.15 g; 89.5 mmol) was added. The resulting
mixture was stirred for 18 h at ambient temperature. Water (90 mL)
was added, and the resulting suspension stirred for 1.5 h. The solid
was isolated by filtration, washing with water (100 mL), and dried
under suction for 20 min. The damp filter cake was slurried in 10%
aqueous citric acid (100 mL) for 1 h. The solid was isolated by
1262
dx.doi.org/10.1021/op2002408 |Org. Process Res. Dev. 2011, 15, 1256–1265

Organic Process Research & Development
ARTICLE
for 90 min. The slurry was heated to reflux and stirred for 14 h. The
slurry was cooled to 20 °C and filtered to remove the byproduct 18.
The filter cake was washed with acetic acid (38 L). The acetic acid
solution was diluted with water (544 L) and was extracted twice with
a mixture of methanol (17 L) and dichloromethane (153 L). The
retained aqueous phase was diluted with 2-methyltetrahydrofuran
(170 L) and cooled to 5 °C. Concentrated aqueous ammonia (35%,
306 L) was added at such rate as to keep the temperature below 30 °C
(about 1 h), followed by a 10-L water line wash. The pH of the
mixture was checked (specification >9, actual pH was 10), then the
mixture was diluted with 2-methyltetrahydrofuran (170 L), and the
temperature was adjusted to 20 °C. The phases were separated, and
the lower aqueous phase was back-extracted with 2-methyltetrahy-
drofuran (170 L). The combined organic extracts were washed with
water (2 ꢁ 136 L) and concentrated under vacuum (around 50
mbar) at 20 °C to approximately 255 L total volume. Acetonitrile
(255 L) was added, and the mixture was concentrated under vacuum
(around 50 mbar) at 20 °C to approximately 255 L total volume.
Acetonitrile (205 L) was added, and the mixture was concentrated
under vacuum (around 50 mbar) at 20 °C to approximately 255 L
total volume. The resulting slurry was diluted with acetonitrile (255 L)
and aged at 20 °C for 4 h. The solid was isolated by filtration, washed
with acetonitrile (2 ꢁ 68 L), and dried at 40 °C under vacuum to give
the title compound 9 as a white solid (20.24 kg; 64%).
N-{2-(Benzyloxy)-5-[(2R)-oxiran-2-yl]phenyl}-N-(methyl-
sulfonyl)methanesulfonamide 23. Potassium carbonate (2.25
g; 16.3 mmol) was added to a solution of bromohydrin 22 (6.0 g;
12.5 mmol) in a mixture of methanol (30 mL) and THF (30 mL),
and the resultant mixture was stirred at ambient temperature for
18 h. The reaction mixture was poured into water (60 mL) and
extracted with propionitrile (2 ꢁ 60 mL). The combined propio-
nitrile layers werewashed with water (100 mL), dried with anhydr-
ous MgSO4, filtered, and concentrated to yield the title compound
23 (4.98 g; ∼100%) as a pale-yellow solid that was used without
further purification. 1H NMR (CDCl3, 400 MHz) δ: 7.54ꢀ7.48
(2H, m), 7.44ꢀ7.40(2H, m), 7.38ꢀ7.34(2H, m), 7.25(1H, d, J =
2.2 Hz), 7.08 (1H, d, J = 8.6 Hz), 5.17 (2H, s), 3.86 (1H, dd, J =
4.1, 2.5 Hz), 3.35 (3H, s), 3.33 (3H, s), 3.15 (1H, dd, J = 5.5, 4.1
Hz), 2.79 (1H, dd, J = 5.5, 2.5 Hz). 13C NMR (100 MHz, CDCl3)
δ: 156.4, 135.5, 131.1, 129.5, 129.1, 128.7, 128.5, 127.8, 123.3,
113.7, 71.3, 51.5, 51.3, 43.7, 43.7.
2-[3-(2-{[(2R)-2-{4-(Benzyloxy)-3-[(methylsulfonyl)amino]-
phenyl}-2-hydroxyethyl]amino}-2-methylpropyl)phenyl]-
N-[(40-hydroxybiphenyl-3-yl)methyl]acetamide 25. A mixture
of amine 9 (500 mg; 1.29 mmol) and epoxide 23 (670 mg; 1.69
mmol) in butyronitrile (2 mL) was heated at reflux for 20 h under
an inert atmosphere. The mixture was cooled to ambient tem-
perature, andchromatographed directlyon silicagel (40 g), eluting
with methanolꢀdichloromethane (1:19 to 1:9) to provide bis-
mesylate 24 (543 mg; 54%) as a waxy oil. 1H NMR (CD3OD, 400
MHz) δ: 7.52 (2H, m), 7.46 (2H, m), 7.42ꢀ7.25 (9H, m), 7.22
(2H, m), 7.17 (1H, m), 7.13 (1H, m), 7.10 (1H, dt, J = 6.6, 1.6
Hz), 6.84 (2H, m), 5.15 (2H, s), 4.75 (1H, dd, J = 8.2, 4.7 Hz),
4.41(2H, s), 3.56(2H, s), 3.33 (3H, s), 3.32(3H, s), 2.88 (2H, m),
2.74 (1H, d, J = 13.1 Hz), 2.68 (1H, d, J = 13.1 Hz), 1.07 (3H, s),
1.05 (3H, s). 13C NMR (CD3OD, 100 MHz) δ: 174.1, 158.3,
157.5, 142.7, 140.4, 138.8, 137.5, 137.4, 136.9, 133.5, 132.5, 131.4,
130.7, 130.4, 130.0, 129.7, 129.5, 129.4, 129.1, 129.1, 128.5, 126.6,
126.4, 126.3, 124.6, 116.8, 114.8, 72.3, 72.1, 56.1, 50.4, 46.8, 44.3,
44.1, 44.1, 44.0, 26.2, 25.8. LCMS: m/z786.36 [M + H]+. Asolution
of sodium hydroxide (500 mg; 12.5 mmol) in water (5 mL) was
added to a solution of 24 (500 mg; 0.64 mmol) in ethanol (5 mL),
and the resulting yellow solution was stirred at ambient temperature
until reaction completion. The mixture was diluted with water
(10 mL) and washed with dichloromethane (10 mL). The aqueous
phase was adjusted to pH 1 with hydrochloric acid and extracted
with propionitrile (2 ꢁ 20 mL). The combined propionitrile
extracts were washed with water, dried with anhydrous MgSO4,
filtered, and concentrated to give the title compound 25 (272 mg;
60%) as a pale-yellow glassy solid. 1H NMR (CD3OD, 400 MHz)
δ: 7.50ꢀ7.46 (3H, m), 7.40ꢀ7.26 (8H, m) 7.20ꢀ7.11 (5H, m),
7.05 (1H, d, J = 8.6 Hz), 7.01 (1H, m), 6.84 (2H, m), 5.18 (2H, s),
4.66 (1H, dd, J = 8.4, 4.5 Hz), 4.41 (2H, s), 3.55 (2H, s), 2.84
(3H, s), 2.82 (1H, m), 2.72 (1H, dd, J = 11.3, 4.5 Hz), 2.65 (1H, d,
J = 13.1 Hz), 2.59 (1H, d, J = 13.1 Hz), 1.01 (3H, s), 0.98 (3H, s). 13C
NMR (CD3OD, 100 MHz) δ: 174.1, 158.4, 151.9, 142.7, 140.4, 139.6,
138.1, 137.5, 136.7, 133.5, 132.4, 130.2, 130.0, 129.8, 129.4, 129.3, 129.1,
129.0, 128.2, 127.5, 126.6, 126.4, 126.3, 125.4, 124.1, 116.8, 114.0, 73.6,
71.8, 54.3, 50.7, 47.5, 44.3, 44.0, 40.0, 27.0, 26.6.
N-{2-(Benzyloxy)-5-[(1R)-2-bromo-1-hydroxyethyl]phenyl}-
N-(methylsulfonyl)methanesulfonamide 20. Bromide 12 (20.0 g;
39.2 mmol) and diisopropylethylamine (24 mL; 138 mmol) were
combined in acetonitrile (100 mL) and cooled to 5 °C. Methane-
sulfonyl chloride (9.0 mL; 118.8 mmol) was added over 10 min, and
the resultant mixture was stirred for about 1 h at 5 °C. Water
(300 mL) was added, and the resultant slurry was stirred for 15 min,
filtered, and dried at 40 °C under vacuum to provide the TBS
protected bis-mesylate 21 (23.3 g; 100%) as a pale-yellow solid.
1H NMR (CDCl3, 400 MHz) δ: 7.53ꢀ7.47 (2H, m), 7.46ꢀ7.33
(5H, m), 7.08 (1H, d, J = 8.6 Hz), 5.16 (2H, s), 4.86 (1H, dd, J = 7.4,
4.7 Hz), 3.51ꢀ3.42 (2H, m), 3.34 (3H, s), 3.33 (3H, s), 0.93 (s, 9H),
0.15 (s, 3H), ꢀ0.03 (s, 3H). 13C NMR (CDCl3, 100MHz) δ:156.2,
135.7, 135.6, 130.3, 129.6, 128.9, 128.7, 128.5, 127.8, 113.3, 74.1, 71.2,
43.7, 39.3, 25.8, 18.3, ꢀ4.7, ꢀ4.9. LCMS: Found m/z 609.12/611.12
[M + NH4]+. TBS-protected bis-mesylate 21 (19.2 g; 32.4 mmol)
was suspended in a mixture of tetrahydrofuran (40 mL) and
methanol (2 mL). Triethylamine trihydrofluoride (9 mL; 55.2
mmol) was added, and the resultant solution was stirred for 30 h at
ambient temperature. The reaction was quenched with aqueous
ammonia (35%, 20 mL), and the product was extracted into ethyl
acetate (2 ꢁ 30 mL). The combined organic phases were washed
with saturated aqueous sodium hydrogen carbonate and water, dried
with anhydrous MgSO4, filtered, and concentrated to dryness. The
residue was slurried in ethyl acetate (40 mL) for 2 h, after which time
the product was isolated by filtration, washing with ethyl acetate
(10 mL) and tert-butyl methyl ether (20 mL). The solid was dried at
40 °C under vacuum for 18 h to give the title compound 22 (11.3 g,
1
73%) as a white solid. H NMR (400 MHz, DMSO-d6) δ: 7.50
(4H, m), 7.42ꢀ7.32 (3H, m), 7.21 (1H, d, J = 8.4 Hz), 5.87 (1H, d,
J = 4.9 Hz), 5.23 (2H, s), 4.80 (1H, m), 3.69 (1H, dd, J = 10.2,
4.3 Hz), 3.59 (1H, dd, J = 10.2, 7.4 Hz), 3.43 (3H, s), 3.42 (3H, s).
13C NMR (100 MHz, DMSO-d6) δ: 155.3, 136.3, 135.5, 130.4,
129.4, 128.4, 128.0, 127.5, 122.3, 113.2, 71.0, 70.1, 43.6, 40.1. LCMS:
Found m/z 495.02/497.02 [M + NH4]+. Anal. Calcd For
C17H20BrN2O6S2: C, 42.68; H, 4.21; N, 2.93; S, 13.41. Found: C,
42.64; H, 4.17; N, 2.93; S, 13.52.
2-[3-(2-{[(2R)-2-{4-(Benzyloxy)-3-[(methylsulfonyl)amino]-
phenyl}-2-hydroxyethyl]amino}-2-methylpropyl)phenyl]-
N-[(40-hydroxybiphenyl-3-yl)methyl]acetamide Fumaric Acid
Salt 25a. Amine 9 (16.5 kg; 42.5 mol), bromide 12 (22.9 kg;
44.6 mol), and sodium hydrogen carbonate (7.1 kg; 85 mol)
were added to n-butyl acetate (165 L), and the resulting slurry
was heated to reflux. After 15 min at reflux, the mixture was
1263
dx.doi.org/10.1021/op2002408 |Org. Process Res. Dev. 2011, 15, 1256–1265

Organic Process Research & Development
ARTICLE
concentrated by distillation to approximately 80 L reaction volume
(120 L was distilled out over 2 h). The resulting slurry was refluxed
for a further 37 h, at which point HPLC analysis showed no more
than 5% 9 remaining, and the batch was progressed. The mixture
was cooled to 50 °C, diluted with ethyl acetate (165 L), cooled to
20 °C, and washed with water (165 L). The retained organic phase
was diluted with methanol (50 L), and triethylamine trihydro-
fluoride (6.8 kg; 42.5 mol) was added, followed by an ethyl acetate
line wash (8 L). The solution was stirred at 20 °C for 3 h, at which
point HPLC analysis indicated complete consumption of silyl
ether 26. A mixture of water (79 L) and concentrated aqueous
ammonia (33 L) was added, followed by a water line wash (20 L),
and the mixture was stirred for 20 min. The phases were separated,
and the retained organic phase was washed with water (66 L) and
then concentrated under vacuum at 30 °C to approximately 75 L
reaction volume. The thick mixture was cooled to 20 °C and diluted
with MEK (248 L). Water (3.3 L) was added, and the mixture was
stirred until complete dissolution occurred (around 30 min).
Fumaric acid (2.5 kg; 21.25 mol) was added; the salt was observed
to oil out of solution. This oily mixture was heated to reflux for 4 h,
during which time it converted into a thick, cream-colored suspen-
sion. Once crystallization had occurred, the batch was refluxed for a
further 1 h, cooled to 23 °C, and aged for 4 h. The solid was isolated
by filtration, washed with MEK (2 ꢁ 165 L), and dried at 40 °C
under vacuum to give the product 25a (28.27 kg; 87%) as an off-
white solid. Mp 129 °C; 1H NMR (400 MHz, DMSO-d6) δ: 8.61
(1H, t, J = 5.9 Hz), 7.56 (2H, dm, J = 7.0 Hz), 7.44ꢀ7.30 (9H, m),
7.26ꢀ7.19 (3H, m), 7.17ꢀ7.06 (4H, m), 6.87 (2H, dm, J =8.8 Hz),
6.55 (1H, s), 5.19 (2H, s), 4.77 (1H, dd, J = 9.0, 2.5 Hz), 4.34 (2H,
d, J = 5.9 Hz), 3.50 (2H, s), 3.02ꢀ2.96 (1H, m), 2.93 (3H, s),
2.88ꢀ2.78 (3H, m), 1.08 (3H, s), 1.07 (3H, s). 13C NMR
(100 MHz, DMSO-d6) δ: 170.2, 169.1, 157.3, 150.8, 140.2,
139.9, 136.8, 136.1, 135.8, 135.6, 131.3, 130.6, 128.7, 128.6, 128.3,
128.1, 127.8, 127.6, 127.6, 127.1, 125.6, 125.2, 124.7, 124.6, 124.3,
124.2, 124.0, 115.7, 112.8, 69.8, 56.0, 53.3, 49.1, 45.2, 44.3, 42.4,
42.2, 24.1, 24.1. LCMS: Found m/z 708.41 [M + H]+. Anal. Calcd
wash (10 L). The resulting mixture was hydrogenated at 20 °C
under 3.5 bar hydrogen for 5 h, at which point hydrogen up-
take had ceased and HPLC analysis confirmed reaction com-
pletion. The slurry was filtered through a Gauthier filter, and the
spent catalyst was washed with a mixture of THF (45 L)
and water (5 L). The resulting solution was filtered through a
1.2 μm filter and concentrated by distillation to around 160 L.
Acetonitrile (50 L) was added, and the mixture was distilled
down to 160 L. This process was repeated until the reflux
temperature exceeded 81 °C (this required eight separate 50-L
acetonitrile charges), and the volume was adjusted back to 160
L. The slurry was cooled to 20 °C at a rate of 0.5 °C/min and
aged for 4 h at 20 °C. The solid was isolated by filtration, washed
with acetonitrile (2 ꢁ 50 L), and dried under vacuum in a tray
drier at 40 °C for 20 h, to give crude 8 (6.48 kg). This crude
material was suspended in a mixture of methanol (58.3 L) and
water (6.5 L), and the resulting slurry was stirred at 50 °C for 2
h. The mixture was cooled to 20 °C at 1 °C/min and then aged for
18 h at 20 °C. The solid was isolated by filtration, washed with a
mixture of methanol (58.3 L) and water (6.5 L), and then dried at
40 °C under vacuum in a tray drier to give 8 (5.65 kg, 70%) as a
white solid. Analytical data were identical to those reported
previously.1,2
Alternative Procedure. A mixture of bromohydrin 12 (10.93 g;
21.2 mmol), amine 9 (7.50 g; 19.3 mmol) and sodium hydrogen
carbonate (9.0 g; 107.1 mmol) in n-butyl acetate (55 mL) was
refluxed under nitrogen for 53 h. The mixture was cooled to
ambient temperature and diluted with water (180 mL) and ethyl
acetate (180 mL). The phases were separated, and the organic
phase was washed successively with aqueous (L)-tartaric acid
(1 M, 55 mL), water (55 mL), a mixture of water and 35%
aqueous ammonia (3:1, 60 mL), and water (55 mL). Palladium on
carbon catalyst (5%, 50% water wet; 1300 mg) was added, and the
resulting mixture was hydrogenated at 60 °C and 4 bar hydrogen
pressure for 24 h. The reaction mixture was removed from the
hydrogenation reactor, and Arbocel (13 g) was added. The resulting
slurry was stirred for 30 min. The mixture was filtered through a pad
of Arbocel, and the catalyst bed was washed with ethyl acetate
(200 mL). The pale-yellow filtrate was concentrated under
reduced pressure to remove the ethyl acetate. Then methanol
(60 mL) was added, and the mixture was concentrated to dryness
under reduced pressure. The resulting viscous, orange-brown oil
was dissolved in methanol (100 mL) and placed in a polypro-
pylene vessel. Ammonium fluoride (2.1 g; 56.7 mmol) was
added, washing with water (20 mL) and methanol (20 mL),
and the resulting solution was stirred at ambient temperature for
65 h. The precipitated solid was isolated by filtration, washed
with a mixture of methanol (80 mL) and water (20 mL), and
dried at 40 °C under vacuum for 4 h. The pale-brown solid was
slurried in a mixture of methanol (67.5 mL) and water (7.5 mL)
at 50 °C for 2 h, and at ambient temperature for 16 h. The
solid was isolated by filtration, washed with methanolꢀwater
(8:2, 2 ꢁ 20 mL), and dried at 40 °C in a vacuum oven for
18 h. The resulting solid was slurried in water (80 mL) at
ambient temperature for 16 h, isolated by filtration, and
washed with water (50 mL) to give 8 (6.01 g; 50%) as an off-
white solid.
For C41H45N3O6S 1/2(C4H4O4): C, 67.43; H, 6.19; N, 5.49; S,
3
4.19. Found: C, 66.96; H, 6.18; N, 5.54; S, 4.07.
N-[(40-Hydroxybiphenyl-3-yl)methyl]-2-[3-(2-{[(2R)-2-hydro-
xy-2-{4-hydroxy-3-[(methylsulfonyl)amino]phenyl}ethyl]-
amino}-2-methylpropyl)phenyl]acetamide 8. A suspension
of fumarate 25a (10.0 kg; 13.1 mol) in a mixture of THF (100 L)
and water (20 L) was stirred at 20 °C while aqueous ammonia
(35%, 10 L) was added, followed by a 10-L water line wash.
The suspension was stirred until a clear solution was obtained
(40 min in this case). The reactor contents were settled, and the
lower aqueous phase was removed. The organic solution was
washed with a solution of NaCl (3.6 kg) in water (26 L). HPLC
analysis of a sample confirmed that the salt break was complete
(NMT 1% fumaric acid remaining). The solution was then
diluted with THF (50 L) and distilled at atmospheric pressure
to approximately 90 L. Additional THF (50 L) was added, and
the mixture was concentrated by distillation to 90 L. At this point
the reflux temperature was >64 °C, indicating complete removal
of water,31 so the solution was cooled to 20 °C. The solution was
filtered through a 16-in. Cuno carbon cartridge (Zeta carbon
C16ME R54SP) at a flow rate of 10ꢀ12 L/min, follo-
wed by a THF wash (30 L) at the same flow rate. The filtered
solution was transferred to a hydrogenation reactor, followed by
a THF wash (60 L). Separately, a slurry of 5% Pd/C catalyst
(type 87 L; 50% water wet, 1.0 kg) in water (10 L) was prepa-
red; this was added to the THF solution, followed by a water
’ AUTHOR INFORMATION
Corresponding Author
pieter.de.koning@pfizer.com
1264
dx.doi.org/10.1021/op2002408 |Org. Process Res. Dev. 2011, 15, 1256–1265

Organic Process Research & Development
ARTICLE
’ ACKNOWLEDGMENT
(14) Colleagues have recently published a similar Ritter/deprotec-
tion reaction sequence to prepare two β3-agonists, see Bradley, P. A.;
Carroll, R. J.; Lecouturier, Y. C.; Moore, R.; Noeureuil, P.; Patel, B.;
Snow, J.; Wheeler, S. Org. Process Res. Dev. 2010, 14, 1326.
(15) Jirgensons, A.; Kauss, V.; Kalvinsh, I.; Gold, M. R. Synthesis
2000, 1709.
(16) It is essential to remove the chloroacetonitrile before the free
base (9) is liberated, as this reacts with chloroacetonitrile during the
solvent exchange to acetonitrile, yielding the corresponding cyano-
methyl adduct. In addition to reducing the yield, this impurity adversely
affects the crystallization of 9.
We thank Rob Bright for his expertise in transferring this
process to the pilot plant and for process safety work; Watcharee
Cooper and Guy Matthews for analytical support; Melissa Birch,
Stephane Dubant, Stuart Field, Steve Fussell, Ben Mathews, Neal
Sach, and Stefan Taylor for reaction, purification, and salt
screening support; Chris Ashcroft, Stewart Eccles, and Imelda
McCarthy for preliminary experimental work; Mike Hawksworth
for process safety studies; John Deering and Trevor Newbury for
hydrogenation support; Stephane Content, David Entwistle,
Nico Fedou, Julian Smith, and our academic consultant, Professor
Steve Ley, for valuable discussions and experimental support;
Helen Barker, Jeremy Clarke, Joe DiBrino, Samantha Farenden,
and Jennifer LaFontaine for their support for and contributions to
this project.
(17) Related epoxides with an additional protecting group on the
sulfonamide are known in the patent literature, e.g. benzyl-protected,
Johnson, M. R.; Hirsh, A. J.; Boucher, R. C.; Zhang, J. Pat. Appl. WO/
2007/146869 A1.
(18) A ‘typical’ reaction HPLC trace (uncorrected) shows approxi-
mately 17% 9, 60% 24, 10% regioisomer, and 8% overalkylated ‘dimeric’
species (the regioisomer and dimers were assigned on the basis of MS data
only and have not been isolated), as well as other low-level impurities.
Similar regioisomeric and dimeric byproducts have been reported for the
coupling of the analogous benzyl-protected epoxide, see reference 17.
(19) While we have only prepared epoxide 23 using the route
described, a shorter route analogous to that used for the synthesis of
the epoxide utilized in the manufacture of formoterol (five steps) should
be feasible, and other, even shorter routes, can be envisaged.
(20) An initial screen of 19 bases in two solvents (DMF, NMP) was
conducted, and then the two best hits from this first screen (NaHCO3
and K2HPO4) were examined in 11 different solvents.
(21) ICH Topic Q3C: Impurities: Guideline for Residual Solvents,
(R4); European Medicines Agency: London, issued February 2009;
http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_
guideline/2009/09/WC500002674.pdf.
’ REFERENCES
(1) Glossop, P. A.; Lane, C. A. L.; Price, D. A.; Bunnage, M. E.;
Lewthwaite, R. A.; James, K.; Brown, A. D.; Yeadon, M.; Perros-Huguet,
C.; Trevethick, M. A.; Clarke, N. P.; Webster, R.; Jones, R. M.; Burrows,
J. L.; Feeder, N.; Taylor, S. C. J.; Spence, F. J. J. Med. Chem. 2010, 53,
6640.
(2) de Koning, P. D.; Gladwell, I. R.; Moses, I. B.; Panesar, M. S.;
Pettman, A. J.; Thomson, N. M. Org. Process. Res. Dev. 2011, DOI:
10.1021/op2001904.
(3) de Koning, P. D.; Gladwell, I. R.; Morrison, N. A.; Moses, I. B.;
Panesar, M. S.; Pettman, A. J.; Thomson, N. M.; Yazbeck, D. R. Org.
Process Res. Dev. 2011, 15, 871.
(4) Other protecting groups were evaluated, but the benzyl and TBS
groups used in the initial synthesis were the preferred choices.
(5) Coe, D. M.; Perciaccante, R.; Procopiou, P. A. Org. Biomol. Chem.
2003, 1, 1106.
(22) Only trace levels of the N-acyl impurity arising from reaction of
amine 9 with n-butyl acetate were observed.
(23) The product mixture was poorly soluble in n-butyl acetate at
ambient temperature.
(24) Chen, C. ꢀK.; Singh, A. K. Org. Process Res. Dev. 2001, 5, 508.
(25) Solvents examined included ethanol, methanol, and THF. It
appears as if the residual reagent and byproducts solubilize 8, hampering
crystallization.
(26) Pilcer, G.; Amighi, K. Int. J. Pharm. 2010, 392, 1.
(27) A range of carbons were evaluated from which Cuno Zeta
carbon R54SP was selected.
(28) We were not able to identify the catalyst poison.
(29) Most other Pd scavengers examined caused some degradation
and/or afforded poor recovery of 8.
(6) Hett, R.; Fang, Q. K.; Gao, Y.; Wald, S. A.; Senanayake, C. H. Org.
Process Res. Dev. 1998, 2, 96.
(7) (a) Baur, F.; Beattie, D.; Beer, D.; Bentley, D.; Bradley, M.;
Bruce, I.; Charlton, S. J.; Cuenoud, B.; Ernst, R.; Fairhurst, R. A.; Faller,
B.; Farr, D.; Keller, T.; Fozard, J. R.; Fullerton, J.; Garman, S.; Hatto, J.;
Hayden, C.; He, H.; Howes, C.; Janus, D.; Jiang, Z.; Lewis, C.; Loeuillet-
Ritzler, F.; Moser, H.; Reilly, J.; Steward, A.; Sykes, D.; Tedaldi, L.;
Trifilieff, A.; Tweed, M.; Watson, S.; Wissler, E.; Wyss, D. J. Med. Chem.
2010, 53, 3675. (b) Zaborenko, N.; Bedore, M. W.; Jamison, T. F.;
Jensen, K. F. Org. Process Res. Dev. 2011, 15, 131.
(8) There is precedent for construction of a tert-butyl amine via imine
formation and addition of MeMgBr (Chung, J. Y. L.; Cvetovich, R. J.;
McLaughlin, M.; Amato, J.; Tsay, F.-R.; Jensen, M.; Weissman, S.; Zewge,
D. J. Org. Chem. 2006, 71, 8602). However, routes to a headgroup amine
were lengthy (2ꢀ3 steps from 1) and those to a suitable ketone (as shown
in Scheme 2) were of comparable length to those for the preparation of 9;
thus, this approach was deemed lower priority and was not investigated.
(9) Several alternative strategies were examined; however, these met with
little success. Examples include asymmetric reduction of an amino ketone (see
Smith, P.; Brodfuehrer, P. R.; Dillon, J. L.; Vemishetti, P. Synth. Commun.
1995, 25, 1093) and asymmetric reduction of an iminoketone (see Hong, Y.;
Gao, Y.; Nie, X.; Zepp, C. M. Tetrahedron Lett. 1994, 35, 5551).
(10) Brough, P.; Pecaut, J.; Rassat, A.; Rey, P. Chem.—Eur. J. 2006,
12, 5134.
(30) The initial water washes are to reduce the level of TFA in order
to minimize off-gassing in the subsequent potassium hydrogen carbo-
nate washes.
(31) The THFꢀwater azeotropic mixture distills at approximately
63 °C, while pure THF has a bp of 65ꢀ67 °C.
(11) We did not attempt to identify any byproducts or conduct a full
mass-balance analysis.
(12) Shokova, E.; Mousoulou, T.; Luzikov, Yu.; Kovalev, V. Synthesis
1997, 1034.
(13) (a) Jiang, X.; Lee, G. T.; Villhauer, E. B.; Prasad, K.; Prashad, M.
Org. Process Res. Dev. 2010, 14, 883. (b) Baum, J. C.; Milne, J. E.; Murry,
J. A.; Thiel, O. R. J. Org. Chem. 2009, 74, 2207. (c) Chang, S.-J. Org.
Process Res. Dev. 1999, 3, 232.
1265
dx.doi.org/10.1021/op2002408 |Org. Process Res. Dev. 2011, 15, 1256–1265