Discovery and development of an efficient, scalable, and robust route to the novel CENP-E inhibitor GSK923295A

Article abstract of DOI:10.1021/op100186c

The discovery and development of an efficient manufacturing route to the CENP-E inhibitor 3-chloro-N-{(1S)-2-[(N,N-dimethylglycyl)amino]-1-[(4-{8-[(1S)- 1-hydroxyethyl]imidazo[1,2-a]pyridin-2-yl}phenyl)methyl]ethyl} -4-[(1-methylethyl)oxy]benzamide (GSK923295A) is described. The existing route to GSK923295A was expensive, nonrobust, used nonideal reagents, and consistently struggled to deliver the API needed for clinical studies. The new synthesis commences from the readily available l-phenylalaninol, which is smoothly converted through to GSK923295A using key Friedel-Crafts acylation as well as selective acylation chemistries. Downstream chemistry to GSK923295A is both high yielding and robust, and the resulting process has been demonstrated first on the kilo scale and subsequently in the pilot plant where 55 kg was successfully prepared. The resulting process is simple, uses cheaper raw materials, is greener in that it avoids using aluminum, tin, and bromination chemistries, and obviates the need for chromatographic purification. Also discussed are the route derived impurities, how they were unambiguously prepared to confirm structure and processing amendments to control their formation, and enhancements to the new process to facilitate future processing.

Full text of DOI:10.1021/op100186c

Organic Process Research & Development 2010, 14, 1254–1263
Discovery and Development of an Efficient, Scalable, and Robust Route to the Novel
CENP-E Inhibitor GSK923295A
Richard Bellingham,† A. Mark Buswell,† Bernie M. Choudary,† Andrew H. Gordon,*,† Steve O. Moore,† Matthew Peterson,‡
Mike Sasse,† Amin Shamji,† and Michael W. J. Urquhart*,†
Synthetic Chemistry, GlaxoSmithKline Pharmaceuticals, Old Powder Mills, Nr. Leigh, Tonbridge, Kent TN11 9AN, U.K., and
Cytokinetics, 280 East Grand AVenue South, San Francisco, California 94080, U.S.A.
Abstract:
causes cell cycle arrest in mitosis with a bipolar mitotic spindle
with misaligned chromosomes, often followed by cell death.3
GSK923295A is a novel and selective inhibitor of CENP-E
motor domain enzymatic activity and is currently in phase I
clinical trials for the treatment of cancer.
Herein is described the limitations of the original supply
route, the new route investigation and subsequent scale-up, as
well as identification and control of some of the key route
impurities.
The discovery and development of an efficient manufacturing
route to the CENP-E inhibitor 3-chloro-N-{(1S)-2-[(N,N-
dimethylglycyl)amino]-1-[(4-{8-[(1S)-1-hydroxyethyl]imida-
zo[1,2-a]pyridin-2-yl}phenyl)methyl]ethyl}-4-[(1-methyleth-
yl)oxy]benzamide (GSK923295A) is described. The existing
route to GSK923295A was expensive, nonrobust, used non-
ideal reagents, and consistently struggled to deliver the API
needed for clinical studies. The new synthesis commences
from the readily available L-phenylalaninol, which is smoothly
Results and Discussion
converted through to GSK923295A using key Friedel-Crafts
acylation as well as selective acylation chemistries. Down-
stream chemistry to GSK923295A is both high yielding and
robust, and the resulting process has been demonstrated first
on the kilo scale and subsequently in the pilot plant where
55 kg was successfully prepared. The resulting process is
simple, uses cheaper raw materials, is greener in that it
avoids using aluminum, tin, and bromination chemistries,
and obviates the need for chromatographic purification. Also
discussed are the route derived impurities, how they were
unambiguously prepared to confirm structure and processing
amendments to control their formation, and enhancements
to the new process to facilitate future processing.
Initial Supply Route to GSK923295A. The existing route
used in outsource campaigns was 10 stages long and delivered
GSK923295A in an overall yield of 4.1% (Scheme 1). The
chemistry involved chemoselective reduction of the expensive
bromophenylalanine 1 to the corresponding alaninol 2 using
lithium aluminum hydride, where careful temperature control
was needed to avoid accompanying reduction of the key
bromide function. The subsequent Mitsunobu coupling of 2 with
phthalimide to give 3 and Stille coupling with the expensive
tin reagent 4 to yield the vinyl ether 5 proceeded with the
concomitant formation of nonideal tin waste (4 kg tin waste
per kg of API). Bromination of 5 provides the halide 6, which
condenses with aminopyridine 74 to secure the benzimidazole
8. Acid-mediated removal of the Boc function in 8 and acylation
with 95 led to the amide 10. Removal of the phthalimide nucleus
with hydrazine and acylation with dimethyl glycine provided
GSK923295A after chromatography and crystallization. Overall,
the initial supply route consistently failed to deliver sufficient
active pharmaceutical ingredient (API) to adequately fund
ongoing clinical studies.
Introduction
Kinesins are a superfamily of motor proteins that use the
energy from ATP hydrolysis to transport cellular cargoes along
microtubules.1 Mitotic kinesins are a functional class of kinesins
essential for mitotic spindle assembly and function during cell
division.2 Perturbation of mitotic kinesin activity results in cell
cycle arrest in mitosis and subsequent cell death. Centromere-
associated protein E (CENP-E) is a mitotic kinesin motor protein
functioning exclusively in mitosis to integrate mitotic spindle
mechanics with mitotic checkpoint signaling. Disruption of
CENP-E function using a variety of methods, including small
molecule inhibition of CENP-E kinesin motor domain function,
Identification of a New Route to GSK923295A 1 via
Phthalimide 11. To address the aforementioned issues, we
began to explore alternative approaches for the synthesis of the
key phenacyl halide functionality. We postulated that such a
functionality could be incorporated using Friedel-Crafts chem-
istry,6 and evaluation of the literature showed that the amino
phthalimide 11, accessible from L-phenylalaninol 12, was a
known compound7 that could serve as a substrate for just such
* To whom correspondence should be addressed. E-mail:
andrewhughgordon@hotmail.com, michael.w.urquhart@gsk.com.
† GlaxoSmithKline Pharmaceuticals.
(3) Wood, K. W.; Cornwell, W. D.; Jackson, J. R. Curr. Opin. Pharmacol.
2001, 1, 370.
‡ Cytokinetics.
(1) (a) Vale, R. D. Cell 2003, 112, 467–480. (b) Miki, H.; Setou, M.;
Kaneshiro, K.; Hirokawa, N. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
7004–7011. (c) Goldstein, L. S.; Philip, A. V. Ann. ReV. Cell DeV.
Biol. 1999, 15, 141.
(4) The aminopyridine 7 was provided by Dr Reddy’s and was derived
from an asymmetric reduction on the corresponding 2-acetyl-2-
aminopyridine.
(5) The acid chloride 9 was commercially available from Minakem but
can also be easily accessed from the readily available ethyl 4-hy-
droxybenzoate in four stages. Please see Supporting Information.
(6) Miyahara, Y. J. Heterocycl. Chem. 1979, 16, 1147.
(2) (a) Bergnes, G.; Brejc, K.; Belmont, L. Curr. Top. Med. Chem. 2005,
5, 127. (b) Sharp, D. J.; Rogers, G. C.; Scholey, J. M. Nature 2000,
407, 41.
1254
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development
10.1021/op100186c  2010 American Chemical Society
Published on Web 08/10/2010

Scheme 1. Initial supply route to GSK923295A
an alternative approach to GSK923295A via the phthalimide
intermediate 11 (Figure 1).
thermally unstable and starts to undergo an intramolecular
cyclization at elevated temperatures (>45 °C) to cyclic carbam-
ate 16.8 This was easily managed by controlling the temperature
of formation and most specifically during the isolation and
drying processes.
A further benefit of this approach was that L-phenylalaninol
12 is readily available since it is derived from a natural amino
acid. The synthesis of the ammonium chloride 11 commenced
by protection of the primary amino function within L-phenyl-
alaninol 12 using Boc anhydride to give the intermediate
carbamate 13 in excellent yield. Conversion to the methane-
sulfonate ester 14 was accomplished again in excellent yield,
and this compound was smoothly converted through to the
differentially protected diamine 15 by reaction with potassium
phthalimide. Finally, acid-mediated hydrolysis of 15 gave 11
in effectively quantitative yield (Scheme 2). For expediency,
the above chemistry was telescoped such that the hydroxy
carbamate 13 was not isolated but was directly converted
through to the methanesulfonate ester 14.
Pleasingly, the ammonium salt 11 underwent an AlCl3-
promoted Friedel-Crafts acylation (Scheme 3) with chloro-
acetyl chloride in excellent conversion and good selectivity in
favor of the desired para product 17, and as expected, the amine
function was effectively removed from the reaction as the
corresponding hydrochloride salt. Interestingly, it was shown
that the minor isomer was not the expected ortho isomer but
rather the meta isomer 18. We postulated that this unusual
isomer pattern was due to the ortho position being too sterically
congested to undergo acylation and that the weak electronic
donation from the alkyl system results in the available meta
position being more electronically favored than originally
expected. The electronic deficiency of the aromatic ring is
The above route to amino phthalimide 11 suffered from one
drawback in that the intermediate methanesulfonate ester 14 is
Figure 1. Retrosynthetic analysis of GSK923295A to L-phenylalaninol 12.
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1255

Scheme 2. Preparation of amino phthalimide 11 from L-phenylalaninol 12
Scheme 3. Friedel-Crafts acylation of 11 and conversion through to phthalimide heterocycle 10
highlighted further by the limited choice of Lewis acids capable
of promoting acylation, and only the group 13 halides of
aluminum and gallium were shown to promote acylation to any
significant extent. All other Lewis acids, including the reactive
lanthanide triflates,9 showed no reaction or decomposition under
forcing conditions. Optimization of the reaction conditions
showed that the reaction performed best when conducted in
neat chloroacetyl chloride (20 equiv). The selectivity of the
reaction improves at lower temperatures, with -5 °C being
selected as optimal providing a para/meta ratio of 4.5/1. No
advantage was gained using bromoacetyl bromide with AlBr3.
Quenching the crude reaction mixture into acidic aqueous
alcohols and subsequent filtration gave a product where the
para/meta ratio was enhanced with respect to the reaction
solution. This key observation demonstrated that the meta
isomer 18 was more soluble than the desired para product 17.
Subsequent development showed the meta isomer 18 could be
solvated exclusively by heating the quenched reaction mixture
to reflux. Subsequent cooling and filtration routinely allowed
isolation of the desired para product 17 with minimal meta
contamination (>50/1 para/meta). Analysis of the filtrate
demonstrated the utility of this process showing >80% PAR
meta product 18.10
The conversion of the phenacyl chloride 17 through to
benzamide 19 was highly chemoselective and resulted in a very
clean and high yielding reaction. No impurities resulting from
reaction of the primary amine of 17 with the phenacyl chloride
functionalities in either 17 or 19 were observed. Reaction of
the phenacyl chloride 19 with the aminopyridine 7 proved to
be a slow transformation. The rate for this reaction could be
markedly enhanced by the addition of the phase transfer catalyst
tetrabutylammonium bromide (TBAB). It was postulated that
this rate enhancement was attributable to the conversion of the
phenacyl chloride to the corresponding bromide, and this is
confirmed by the formation of an intermediate during this
reaction as visible by HPLC. An interesting byproduct was
formed during this chemistry at low levels (up to 0.5% PAR)
and was confirmed to be common to both the early supply route
and the new route. LCMS confirmed the structure as the
brominated pyridyl imidazole 20, and this was confirmed
unambiguously through synthesis. Treatment of imidazole 10
with N-bromo succinimide (NBS) in DMF resulted in an
instantaneous conversion to the corresponding brominated
analogue 20 in quantitative yield. This reactivity of pyridyl
(7) Karavoltsos, M.; Mourtas, S.; Gatos, D.; Barlos, K. J. Pept. Sci. 2002,
8, 615.
(10) Protecting the primary amine function as an amide prior to the Friedel
Crafts chemistry was evaluated, and the benzoyl and trifluoroacetamide
groups showed particular promise. Although slightly higher selectivities
for the para acylation were observed with these substrates, the
requirement for a protection and deprotection strategy made this
approach less desirable than simply using the ammonium salt.
(8) See supporting information for DSC, TGA and ARC thermal safety
data.
(9) (a) Dzudza, A.; Marks, T. J. J. Org. Chem. 2008, 73, 4004. (b)
Repichet, S.; Le Roux, C.; Dubac, J.; Desmurs, J.-R. Eur. J. Org.
Chem. 1998, 2743.
1256
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

Scheme 4. Conversion of phthalimide 10 through to GSK923295A
imidazoles toward NBS is known,11 but the exact mechanism
for oxidation of the bromide ion to a source of electrophilic
bromine within the reaction medium is currently unknown.
The phthalimide 10 was smoothly converted to the amine
21 using hydrazine hydrate in denatured ethanol. It was
important to leave this reaction for an extended period of time
to ensure all of the initially formed N-aminophthalimide
rearranges to the more thermodynamically stable phthaloylhy-
drazide,12 thereby rendering the reaction irreversible. It was
observed that minor nonbasic impurities could readily be
removed from this product by incorporating an extractive
workup of 21 into aqueous lactic acid and subsequent basifi-
cation of the aqueous phase, extraction, and isolation to give
21 with a typical purity of 98% PAR. Temperature control of
these reaction conditions was needed to minimize a back ground
“trans-amidation” process whereby 21 could rearrange to the
isomeric secondary amine 22. Under the reaction conditions
used, the level of 22 was always maintained below 0.2% PAR.
To improve the efficiency of the route further, a higher yielding
coupling process was sought for the final stage conversion of
amine 21 to GSK923295A. Experience of activating the acid
as the imidazole ester for amide synthesis has been gained “in-
house”, and this approach has also been successfully used within
the literature with dimethyl glycine.13 Following this precedent,
the preparation of GSK923295A was accomplished in excellent
yield by first preparing an imidazole ester of dimethyl glycine
23 by reaction with carbonyl diimidazole (CDI), and adding 1
equiv of this reagent to amine 21 (Scheme 4). Recrystallization
of GSK923295A from propan-2-ol delivered clinical grade
material in 85% yield.
This route to GSK923295A was scaled up to the 20 L level
where approximately 1 kg of API was prepared in two batches
and was of excellent purity.
During the release analysis of GSK923295A, one new
impurity was identified which was present at the qualification
limit and had the same HPLC retention time as the known
acetamide impurity 24. Structural elucidation of this new
impurity was paramount in order to facilitate the release process.
LCMS suggested and accurate mass highlighted a likely
molecular formula (C32H34ClN5O5). It had also been observed
that this impurity had increased during the azeotropic drying
process of the API solution prior to addition of antisolvent. A
tentative structure 25 was proposed based on the possible
presence of sarcosine within the commercial dimethyl glycine
which would form via the intermediate 26 (Scheme 5). The
structure of this impurity was confirmed through synthesis.14
Analysis confirmed hydantoin 25 as the new impurity, and
1
retrospective H NMR analysis of the dimethyl glycine used
within the scale-up campaign showed that levels of sarcosine
were indeed present at 0.3%, and the level of sarcosine within
dimethyl glycine would need to be controlled through appropri-
ate specification.
As a consequence of the hazardous nature of hydrazine
hydrate and particularly with a view to further scale-up,
alternative conditions were screened to remove the phthalimide
nucleus. Although methylamine15 and ethylene diamine16 were
identified as potential alternatives for this reaction, the quality
of product 21 derived from these conditions was less pure than
that from the hydrazine process. It was likely that further
development would have led to either of these procedures being
suitable for replacing hydrazine hydrate; however, with the tight
delivery timelines for the project and implementation of suitable
engineering within the pilot plant, it was decided to continue
using hydrazine hydrate for this transformation.
(11) (a) Salgado-Zamora, H.; Velazquez, M.; Mejia, D.; Campos-Aldrete,
M. E.; Jimenez, R.; Cervantes, H. Heterocycl. Commun. 2008, 14,
27. (b) Godovikova, S. N.; Gol’dfarb, Ya. L. Bulletin of the Academy
of Sciences of the USSR, DiVision of Chemical Science (English
Translation) 1965, 1391.
(12) (a) Drew, D. K.; Hatt, H. H., J. Chem. Soc., 1937, Abstracts: 16. (b)
Chattaway, F. D.; Tesh, W. J. Chem. Soc. 1920, 117, 711. (c) Hearn,
M. J.; Prisch, S. B. Org. Prep. Proc. Int. 1981, 87, 421.
(14) See supporting information for details.
(15) (a) Khan, M. N.; Ohayagha, J. E. React. Kinet. Catal. L. 1996, 58,
97.
(13) (a) Gerecke, M.; Kyburz, E.; Borer, R.; Gassner, W. Heterocycles
1994, 39, 693. (b) Johnson, D. W. J. Mass Spectrom. 2001, 36, 277.
(16) Bauer, J.; Rademann, J. Tetrahedron Lett. 2003, 44, 5019.
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1257

Scheme 5. Elucidation and synthesis of the hydantoin impurity 24
The process to GSK923295A was transferred to pilot plant,
and a total of 55 kg was prepared. On increased scale, however,
the reaction of phenacyl chloride 19 with aminopyridine 7
resulted in very impure product 10 (85% assay) as a conse-
quence of the semibatch addition mode used. The process
involved stirring the phenacyl chloride 19 with sodium bicar-
bonate in acetonitrile, heating to the reaction temperature, and
subsequently adding a solution of the aminopyridine 7. Under
these conditions, the phenacyl chloride 19 proves to be unstable
and undergoes base-mediated self-condensation, and this was
exacerbated on plant by the prolonged heating and addition
times. Two of the main byproducts were identified by spec-
troscopy as diamide 31 and cyclopropane 32, and their structures
were unambiguously confirmed through synthesis (Scheme 6).
This reactivity of phenacyl halides to base is well-known,17 and
this undesired reactivity is easily remedied by returning to a
formal batch process. The incorporation of an acetonitrile slurry
on the pyridyl imidazole 10 product resulted in a significant
purification to 95% assay. The synthesis of the two key
impurities 31 and 32 present in this large-scale batch of 10 was
required to deliver valuable reference materials. The direct
approach of degrading the phenacyl chloride using base led to
a multicomponent mixture from which it was difficult to isolate
the desired products. A retrosynthetic analysis suggested that
these two compounds were probably accessed via the same
enedione intermediate 30, and it was decided to prepare this
compound from a Wittig reaction between the carboxaldehyde
29 and the phosphonium salt 27. The preparation of Wittig
reagents from chloroacetyl benzene is known;18 similarly,
reaction of the phenacyl chloride 19 with triphenyl phosphine
gave the phosphonium salt 27 in excellent yield. The carbox-
aldehyde 29 was also accessed from the phenacyl chloride 19
in two steps. Initial attempts to hydrolyze 19 led to significant
byproducts forming through aldol-type condensations, or oxida-
tions, of the product 28, which proved a somewhat sensitive
entity. We developed a method for the smooth and particularly
mild conversion of 19 to the hydroxy ketone 28 using potassium
formate to form an intermediate formate ester. Ester hydrolysis
was achieved in situ under the aqueous conditions used for the
reaction to give the hydroxy ketone 28 in excellent yield. The
process was conducted under nitrogen to prevent undesired
oxidative byproducts. There are reports in the literature
whereby phenacyl chlorides can be converted to the
corresponding hydroxy ketones using sodium19 or potas-
sium acetate,20 but in our hands, the reactions were less
clean, required acid hydrolysis, and consequently led to
the undesired aldol chemistry unless very carefully
controlled. Oxidation of the hydroxy ketone to the glyoxal
29 was accomplished in excellent yield using copper
acetate.21 Workup for this reaction proved critical, and
filtration of inorganic residues followed by removal of
solvent prior to extraction proved necessary to minimize
polymer formation. Wittig reaction of the phosphonium
salt 27 with glyoxal 29 using triethylamine as base22 gave
the target diketone 30 in near quantitative yield. Conden-
sation of diketone 30 with the aminopyridine 7 led to the
ketone 31 in very good yield, and in addition, the base-
mediated reaction of enedione 30 with phenacyl chloride
19 gave the cyclopropane 32, also in excellent yield. Both
ketone 31 and cyclopropane 32 were confirmed to be the
impurities present within the crude pyridyl imidazole 10.
Following the successful pilot plant campaign, attention was
focused on some of the processing concerns that would prove
troublesome for future scale-up campaigns. These concerns
included: (1) The methanesulfonate ester 14 was shown to be
mutagenic and was isolated in the current process. (2) Aceto-
nitrile is a class A volatile organic compound and was used in
two stages to prepare compounds 19 and 10. (3) Aminopyridine
7 is a highly viscous oil that required solvation to handle. (4)
The coupling of amine 21 with the imidazole ester 23 requires
titration of the two reagents.
To address the occupational hygiene concerns of
isolating the mutagenic methanesulfonate 14, an alternative
(19) Miyauchi, H.; Nakamura, T.; Ohashi, N. Bull. Chem. Soc. Jpn. 1996,
69, 2625.
(20) Tozuka, H.; Ota, M.; Kofujita, H.; Takahashi, K. J.Wood Sci. 2005,
51, 48.
(17) (a) Saba, A. J. Chem. Res. 1990, 9, 2227. (b) Pasto, D. J.; Garves, K.;
Sevenair, J. P. J. Org. Chem. 1968, 33, 2975.
(21) (a) Nef, Justus Liebigs Ann. Chem. 1904, 335, 82. (b) Evans, American
Chemical Journal 1906, 35, 124.
(18) Uziel, J.; Riegel, N.; Aka, S.; Figuiere, P.; Juge, S. Tetrahedron Lett.
1997, 38, 3405.
(22) Konno, T.; Takehana, T.; Mishima, M.; Ishihara, T. J. Org. Chem.,
2006, 71, 3545.
1258
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

Scheme 6. Preparation of impurities derived from the synthesis of 10
route to the amino phthalimide 11 was investigated from
L-phenylalaninol 12. Activation of the alcohol 13 in situ
utilizing the well-recognized Mitsunobu approach23 pre-
sented an alternative and potentially simpler one-pot
procedure toward alkylation. Gratifyingly, the DIAD/PPh3-
mediated displacement with phthalimide readily provided
the desired 15. There was no competing intramolecular
cyclization, and the need for a caustic slurry to remove
undesired phthalimide was obviated, thereby saving on
processing time. Moreover, the generated triphenylphos-
phine oxide could be efficiently solubilized by addition
of alcoholic solvents while also acting as an antisolvent
for the desired product, thereby providing a simplified
workup procedure. This protocol was easily incorporated
with the tert-butyl carbamate protection and when per-
formed in toluene allowed for removal of the tert-butyl
alcohol byproduct through distillation that would otherwise
consume an equivalent of DIAD and triphenylphosphine
by dehydration. A further benefit of this strategy was that
tert-butyl carbamate cleavage could also be performed in
thus providing a simple three-stage one-pot telescoped
procedure for the synthesis of 11 in 86% yield over the
three steps (Scheme 7).
To avoid the use of acetonitrile, a solvent screen was
conducted for the preparation of both intermediates 10 and 19.
It was quickly demonstrated that THF was a good alternative
with comparable yields of product being obtained in both cases.
Handling concerns associated with the use of the amino-
pyridine 7 were readily addressed by preparing the correspond-
ing hydrochloride salt 33, which proved to be a highly granular
and free-flowing solid. The preparation was facile and very high
yielding and involved treating an ethyl acetate solution of 7
with 4-6 N hydrochloric acid in propan-2-ol. This material
performed very well in the reaction to prepare pyridyl imidazole
10 from phenacyl chloride 19 and simply required the addition
of an extra equivalent of base.
Finally, the procedure to prepare GSK923295A had required
effective titration of the imidazole ester 23 with the amine 21
to avoid undesired secondary O-acylation. This had not proved
an issue in the pilot plant but did require an element of control,
and as such, it was desirable to avoid complicated processing.
Work to prepare the hydantoin impurity 25 from amine 21
(Scheme 5) had demonstrated that reaction of 21 with chloro-
the same reaction by the addition of methanol/conc HCl(aq)
,
(23) Kumara Swamy, K. C.; Bhuvan Kumar, N. N.; Balaraman, E.; Pavan
Kumar, K. V. P. Chem. ReV. 2009, 109, 2551.
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1259

Scheme 7. Telescoped conversion of alcohol 12 through to amino phthalimide 11 using a Mitsunobu approach
Scheme 8. Modified Schotten-Baumann conditions for the preparation of GSK923295A
acetyl chloride leads to an intermediate chloroacetamide that
can react with amines. Extension of this chemistry by quenching
with dimethylamine did deliver GSK923295A, but the byprod-
ucts derived from overacylation were still apparent at low levels.
To circumvent this overacylation, Schotten-Baumann chemistry
was developed, where by reaction of 21 with chloroacetyl
chloride under the biphasic conditions of water and 2-meth-
yltetrahydrofuran, using an appropriate buffer, led to a highly
N-selective acylation.24 Subsequent treatment with aqueous
dimethylamine led to an excellent conversion to GSK923295A,
delivering a simpler and more robust process that would be
used in any future campaigns (Scheme 8).
propoxy-3-chloro benzoyl chloride and 2-amino-3-(1S)-1-hy-
droxyethylpyridine were obtained commercially from custom
synthesis suppliers. The experimental conditions below refer
to the procedures used for the multikilogram pilot plant
campaign. For a key to 1H NMR ring assignments, see
Supporting Information.
(2S)-2-({[(1,1-Dimethylethyl)oxy]carbonyl}amino)-3-phe-
nylpropyl methanesulfonate (14): L-Phenylalaninol 12 (500
g, 3.31 mol) was suspended in ethyl acetate (6.67 L) and treated
with di(tert)-butyl dicarbonate (794 g, 3.64 mol). On complete
reaction (1.5 h), the reaction mixture was cooled to 0 °C and
treated with triethylamine (608 mL, 4.37 mol). Methanesulfonyl
chloride (281 mL, 414 g, 3.64 mol) was added dropwise,
maintaining an internal temperature of below 4 °C at all times.
Water (4.17 L) was added, and the ethyl acetate was removed
by distillation under vacuum. The product was collected by
filtration under vacuum, washed with water (4.0 L), and dried
under vacuum to yield the title compound (1.05 kg, 95.9%) as
a colorless solid: 1H NMR (400 MHz, CDCl3) δH 1.40 (9H, s,
(CH3)3), 2.69-2.95 (2H, m, CH2Ph), 3.01 (3H, s, SO2CH3),
4.05-4.13 (2H, m, CH2O), 4.20-4.29 (1H, m, CHNH), 4.73
(1H, br d, CHNH), 7.18-7.38 (5H, m, CH Ar); 13C NMR (100
MHz, CDCl3) δC 28.3, 37.3, 50.9, 69.9, 80.0, 127.0, 128.8,
129.3, 136.7 and 155.2; HRMS (ESI+) m/z calcd for [M + Na]+
C15H23NO5SNa 352.1195, found 352.1182.
1,1-Dimethylethyl [(1S)-2-(1,3-dioxo-1,3-dihydro-2H-isoin-
dol-2-yl)-1-(phenylmethyl)ethyl]carbamate (15): Potassium
phthalimide (646.3 g, 3.49 mol) was combined with (2S)-2-
({[(1,1-dimethylethyl)oxy]carbonyl}amino)-3-phenylpropyl meth-
anesulfonate 14 (1.05 kg, 3.17 mol), tetrabutylammonium
bromide (133 g, 0.41 mol), and isopropyl acetate (8.36 L). This
mixture was warmed to 35 °C and stirred for 23 h. Isohexane
(8.36 L) was added, and the reaction mixture was cooled to 0
°C. The crude product was isolated by filtration, washed with
a mixture of isohexane and isopropyl acetate (1:1, 3 L), and
dried under vacuum. The crude solid was suspended in 0.15
M NaOH (10.0 L) and stirred at room temperature for 1 h. The
Summary
To conclude, a robust and scalable route to GSK923295A
has been discovered and developed which is both shorter and
higher yielding than the original supply route (33.6 versus
4.1%). The route exemplifies highly chemoselective Friedel-
Crafts chemistry, which obviates the need for both costly
brominated starting materials and nongreen tin chemistry. An
excellent understanding of the main route derived impurities
has also been demonstrated through their unambiguous syn-
thesis, which in itself demonstrates some interesting and
versatile synthetic methods but also leads to a greater under-
standing of the chemical process. Additionally, the route has
been further developed to design out complexity such that future
campaigns will have less isolations, avoid handling of mutagenic
materials, and have inherently more robust chemistry. It is
anticipated that the overall yield will increase to 38% for the
fully developed process.
Experimental Section
General. Commercially available reagents were used without
further purification. Reactions requiring anhydrous conditions
were performed under an atmosphere of dry nitrogen. 4-Iso-
(24) Brown Ripin, D. H.; Vetelino, M. Synlett. 2003, 15, 2353.
1260
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

mixture was isolated by filtration and the cake washed with
water (3 L). The product was dried under vacuum to yield the
title compound (1.05 kg, 87.1%) as a colorless solid: 1H NMR
(400 MHz, CDCl3) δH 1.22 (9H, s, (CH3)3), 2.72-2.95 (2H,
m, CH2Ph), 3.62-3.79 (2H, m, CH2NPhthal), 4.20-4.35 (1H,
m, CHNH), 4.64 (1H, br d, CHNH), 7.18-7.38, 7.65-7.73
and 7.80-7.88 (9H, m, CH Ar); 13C NMR (100 MHz, CDCl3)
δC 28.1, 39.1, 41.8, 50.3, 79.3, 123.3, 126.7, 128.6, 129.2, 132.0,
133.9, 136.9, 155.5 and 168.5; HRMS (ESI+) m/z calcd for
[M + Na]+ C22H24N2O4Na 403.1634, found 403.1615.
2-[(2S)-2-Amino-3-phenylpropyl]-1H-isoindole-1,3(2H)-
dione hydrochloride (11). Pilot Plant Procedure: To a warmed
(60 °C) suspension of 1,1-dimethylethyl[(1S)-2-(1,3-dioxo-1,3-
dihydro-2H-isoindol-2-yl)-1-(phenylmethyl)ethyl]carbamate 15
(1.05 kg, 2.76 mol) in methanol (10.5 L) was added concen-
trated hydrochloric acid (1.22 L, 1.44 kg, 14.64 mol) over 1 h
(caution: gas evolution). The resulting slurry was stirred at
60 °C for 2.25 h. The mixture was cooled to 0 °C, the slurry
isolated by filtration, and the cake washed with propan-2-ol (2.4
L). The solid was dried under vacuum to yield the title
compound (0.786 kg, 89.8%) as a colorless solid: 1H NMR (400
MHz, DMSO-d6) δH 2.86 (1H, dd, J ) 14.0 and 4.0, CH2aPh),
3.20 (1H, dd, J ) 14.0 and 4.0, CH2bPh), 3.64 (1H, dd, J )13.5
and 4.0, CH2aNPhthal), 3.74 (1H, m, CHNH), 3.75 (1H, dd, J
)13.5 and 4.0, CH2bNPhthal), 3.62-3.79 (2H, m, CH2NPhthal),
7.18-7.28, 7.30-7.35 and 7.80-7.88 (9H, 3 × m, CH Ar),
8.41 (3H, br s, NH3); 13C NMR (100 MHz, DMSO-d6) δC 36.5,
39.0, 50.3, 122.9, 126.7, 128.5, 129.0, 131.7, 134.2, 136.2 and
168.0; HRMS (ESI+) m/z calcd for [M - Cl]+ C17H17N2O2
281.1290, found 281.1288.
Optimized Telescoped Procedure to Compound 11: L-Phe-
nylalaninol 12 (15.0 g, 0.099 mol) was suspended in toluene
(90 mL) with mechanical stirring, and the mixture was treated
with di-tert-butyl dicarbonate (22.7 g, 0.104 mol). The reaction
mixture was warmed to 40 °C and stirred until all solids
dissolved (1.5 h). The reaction mixture was heated and
distillation initiated and continued until all tert-butyl alcohol
was removed as determined by 1H NMR analysis. The reaction
mixture was cooled to 70 °C, diluted with toluene (141 mL),
and further cooled to 42 °C. The reaction mixture was treated
with triphenyl phosphine (35.52 g, 0.135 mol) and phthalimide
(15.30 g, 0.104 mol), and DIAD (24.6 mL, 25.1 g, 0.124 mol)
was added, maintaining a temperature between 42 and 52 °C.
After 10 min, methanol (350 mL) was added and the reaction
warmed to 60 °C. Concentrated hydrochloric acid (35 mL) was
added dropwise over 1 h, and the resulting suspension was
stirred for a further hour before cooling to 5 °C (caution: gas
evolution). The product was isolated by filtration and the cake
washed with propan-2-ol (45 mL). The solid was dried under
vacuum to yield the title compound (27.1 g, 86%) as a colorless
solid that was spectroscopically identical to that reported above.
2-{(2S)-2-Amino-3-[4-(chloroacetyl)phenyl]propyl}-1H-
isoindole-1,3(2H)-dione (17): 2-[(2S)-2-Amino-3-phenylpro-
pyl]-1H-isoindole-1,3(2H)-dione hydrochloride 11 (0.785 kg,
2.48 mol) was suspended in chloroacetyl chloride (3.95 L, 5.6
kg, 49.6 mol) and cooled to -5 °C with mechanical stirring.
This suspension was then treated with aluminum chloride (1.65
kg, 12.4 mol). The resulting mixture was stirred at -5 °C for
22 h, after which the reaction mixture was carefully added to
methanol (15.7 L), maintaining a temperature below 35 °C. The
resulting stirred mixture was warmed to 50 °C and treated with
3.5 M hydrochloric acid (3.9 L). The mixture was heated further
to 65 °C, stirred at this temperature for 18 h, then cooled to
ambient. The crude product was collected by filtration and
washed with methanolic hydrogen chloride comprising metha-
nol (2.35 L) and 3.5 M hydrochloric acid (0.785 L). The crude
product was purified further by suspending in a mixture of
methanol (7.85 L) and 3.5 M hydrochloric acid (3.14 L) and
warmed to 65 °C. The product was then isolated by filtration,
washed with methanol (1.57 L), and dried under vacuum to
yield the title compound (0.703 kg, 72.1%) as a colorless solid:
1H NMR (400 MHz, DMSO-d6) δH 2.98 (1H, dd, J ) 14.0
and 9.0, CH2aPh), 3.28 (1H, dd, J ) 14.0 and 4.0, CH2bPh),
3.67 (1H, dd, J )13.3 and 4.0, CH2aNPhthal), 3.74-3.89 (2H,
m, CHNH and CH2bNPhthal), 5.15 (2H, s, CH2Cl), 7.52 (2H,
d, J ) 8.0, CH Ar Phthal), 7.81 (5H, m, CH Ar), 7.92 (2H, d,
J ) 8.0, CH Ar Phthal), 8.46 (3H, br s, NH3); 13C NMR (100
MHz, DMSO-d6) δC 36.4, 39.0, 47.6, 50.3, 122.9, 128.6, 129.5,
131.8, 132.8, 134.2, 142.6, 167.9 and 191.1; HRMS (ESI+) m/z
calcd for [M - Cl]+ C19H18N2O3 357.1006, found 357.0992.
Data for meta isomer 2-{(2S)-2-amino-3-[2-(chloroacetyl)-
phenyl]propyl}-1H-isoindole-1,3(2H)-dione (18): 1H NMR (400
MHz, DMSO-d6) δH 3.02 (1H, dd, J ) 14.0 and 8.5, CH2aPh),
3.30 (1H, dd, J ) 14.0 and 4.5, CH2bPh), 3.72 (1H, dd, J )14.0
and 4.5, CH2aNPhthal), 3.74-3.95 (2H, m, CHNH and
CH2bNPhthal), 5.20 (2H, d, J ) 3.0, CH2Cl), 7.50 (1H, dd, J
) 7.5 and 7.5, CH Ar Phthal), 7.68 (1H, d, J ) 7.5, CH Ar),
7.80 (5H, m, CH Ar), 7.95 (1H, s, CH Ar), 8.60 (3H, br s,
NH3); 13C NMR (100 MHz, DMSO-d6) δC 36.1, 39.0, 47.8,
50.2, 122.9, 126.9, 129.0, 129.1, 131.8, 134.2, 134.5, 134.6,
137.0, 168.0 and 191.4; HRMS (ESI+) m/z calcd for [M - Cl]+
C19H18N2O3 357.1006, found 357.0993.
3-Chloro-N-{(1S)-2-[4-(chloroacetyl)phenyl]-1-[(1,3-dioxo-
1,3-dihydro-2H-isoindol-2-yl)methyl]ethyl}-4-[(1-methyleth-
yl)oxy]benzamide (19): 2-{(2S)-2-Amino-3-[4-(chloroacetyl)phe-
nyl]propyl}-1H-isoindole-1,3(2H)-dione 17 (0.70 kg, 1.78 mol)
was suspended in acetonitrile (9.1 L) and cooled to 5 °C with
stirring. Diisopropylethylamine (0.78 L, 0.575 kg, 4.45 mol)
was added dropwise to the stirred mixture followed by a solution
of 3-chloro-4-[(1-methylethyl)oxy]benzoyl chloride 9 (0.441 kg,
1.89 mol) in acetonitrile (1.3 L), maintaining an addition
temperature below 8 °C. The mixture was stirred at 5-8 °C
for 1 h. Water (10.5 L) was added and the mixture stirred at 5
°C for 30 min. The product was isolated by filtration washing
with a mixture of acetonitrile (1.75 L) and water (3.0 L). The
product was dried under vacuum to yield the title compound
1
(0.962 kg, 97.6%) as a colorless solid: H NMR (400 MHz,
CDCl3) δH 1.42 (6H, d, J ) 6.0, (CH3)2), 3.05 (1H, dd, J )
14.0 and 7.0, CH2aPh), 3.15 (1H, dd, J ) 14.0 and 7.0, CH2bPh),
3.90 (2H, m, CH2NPhthal), 4.60-4.75 (4H, m, CHNH,
(CH3)2CHO and CH2Cl), 6.83 (1H, d, J ) 7.7, NH), 6.96 (1H,
d, J ) 7.5, H-b), 7.42 (2H, d, J ) 8.0, H-d), 7.58 (1H, dd, J )
7.5 and 2.0, H-c), 7.70-7.79 (3H, m, H-b and H-g), 7.82-7.88
(2H, m, H-f), 7.93 (2H, d, J ) 8.0, H-e); 13C NMR (100 MHz,
CDCl3) δC 21.9, 38.8, 41.0, 46.1, 50.8, 71.9, 114.1, 123.5, 123.8,
126.6, 126.7, 128.8, 129.5, 129.7, 131.7, 132.8, 134.3, 144.0,
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1261

156.3, 165.9, 168.8 and 190.7; HRMS (ESI+) m/z calcd for
[M + H]+ C29H27N2O5Cl2 553.1297, found 553.1284.
16 h at this temperature then quenched at 40 °C with water
(4.34 L). The layers were separated, and the organic layer was
washed with water (4.34 L) and the product extracted into an
aqueous solution of 10% lactic acid (3.30 L). The acidic aqueous
layer is washed with TBME (2.6 L) and then basified to pH10
with 4 M NaOH solution (1.22 L) to liberate free amine. The
intermediate 21 was then finally extracted into 2-Me THF (3.47
L), washed with brine (3.47 L) and the solution used directly
in the next stage.
Pilot-plant Procedure. N,N′-Dimethylamino glycine (190.6
g, 1.84 mol) is suspended in acetonitrile (0.762 L) and treated
with carbonyl diimidazole (286.9 g, 1.76 mol) at room tem-
perature. This solution was stirred at room temperature for 30
min and used in the next step after analysis.
2-Amino-3-[(1S)-1-hydroxyethyl]pyridinium chloride (33):
(1S)-1-(2-Amino-3-pyridinyl)ethanol 7 (25 g, 0.181 mol) was
dissolved in ethyl acetate (200 mL) and added dropwise to a
cooled stirred 4-6 N solution of hydrogen chloride in propan-
2-ol (50 mL, 0.2 mol) over 20 min, maintaining an addition
temperature of -5 to 2 °C at all times. The resulting suspension
was stirred at <5 °C for 30 min, and the resulting solid was
collected by filtration, washed with cold ethyl acetate (2 × 25
mL), and pulled dry to give the title compound as a cream solid
(29.9 g, 94.6%): 1H NMR (400 MHz, DMSO-d6) δH 1.32 (3H,
d, J ) 6.5, CH3), 4.88 (1H, q, J ) 6.5, CHCH3), 5.72 (1H, br
s, OH), 6.93 (1H, dd, J ) 7.0 and 6.5, CHAr), 7.90-8.00 (4H,
m, 2 × CHAr and NH2), 14.18 (1H, br s, ArNHCl);13C NMR
(100 MHz, CDCl3) δC 22.5, 63.9, 112.3, 130.4, 133.9, 139.3,
and 151.8; HRMS (ESI+) m/z calcd for [M - Cl]+ C7H11N2O
139.0871, found 139.0867.
Crude 21 (680 g - assay), as a solution in 2-Me THF from
above was stirred at 22 °C. The solution of acid imidate 23
was added dropwise. On complete addition, the reaction was
stirred at room temperature for 30 min before quenching by
the addition of water (5.77 L). The layers were separated and
the aqueous extracted with a further potion of 2-Me THF (1.5
L). The organic layers were combined and distilled under dean
and stark conditions until <4% w/w water was present by Karl
Fischer analysis. The solution was finally distilled to 4.7 L total
volume, cooled to 50 °C and treated with TBME (9.5 L). The
mixture was stirred at 50 °C for 1 h before cooling to 10 °C
and stirring for a further 1 h. The product was isolated by
filtration washing with of 2-Me THF/TBME (1/2, 3.0 L). The
cake was dried under vacuum to yield the title compound (691
g, 87.0%) as a colorless solid. Recystallization from propan-
2-ol (13.6 L) secured clinical grade material (86%). 1H NMR
(400 MHz, CD3OD) δH 1.34 (6H, d, J ) 6.0, (CH3)2), 1.59
(3H, d, J ) 7.0, CH3CH), 2.21 (6H, s, N(CH3)2), 2.87-3.01
(4H, m, CH2Ph and CH2N(CH3)2), 3.49 (2H, m, CH2NPhthal),
4.50 (1H, m, CHNH), 4.70 (1H, m, (CH3)2CHO)), 5.49 (1H,
q, J ) 7.0, CHOH), 6.88 (1H, t, J ) 7.0, H-j), 7.08 (1H, d, J
) 7.5, H-b), 7.33-7.37 (3H, m, H-k and H-d), 7.63 (1H, dd,
J ) 7.5 and 2.0, H-c), 7.78 (1H, s, H-a), 7.83 (2H, d, J ) 7.0,
H-e), 8.09 (1H, m, H-h), 8.27 (1H, d, J ) 8.0, H-i); 13C NMR
(100 MHz, CD3OD) δC 22.2, 24.1, 39.3, 43.8, 46.1, 53.0, 63.7,
66.2, 73.0, 110.4, 113.8, 115.3, 121.2, 124.5, 126.1, 127.5,
128.4, 128.5, 130.6, 130.7, 133.3, 136.0, 139.4, 145.1, 146.1,
157.6, 168.5 and 173.6; HRMS (ESI+) m/z calculated for
[M+H]+ C32H39N5O4Cl 592.2691, found 592.2684.
AlternatiVe Procedure. Intermediate 21 (0.0471 mol) as a
cooled (10 °C) solution in 2-Me THF (300 mL) and water (300
mL) was treated with dipotassium hydrogen phosphate (16.4
g, 0.0942 mol) followed by chloroacetyl chloride (4.85 mL,
7.18 g, 0.0636 mol) dropwise. The reaction mixture was then
treated with an aqueous solution of dimethylamine (40% w/v,
59.6 mL, 0.0471 mol) and heated to 40 °C. After stirring at for
4 h the solution was cooled to room temperature and the layers
separated. The organic layer was further washed with water
(250 mL) and resulting organic layer dried by dean and stark
distillation until <2% w/w water was present as determined by
Karl Fischer analysis. The reaction mixture was concentrated
by distillation to 7 vols total volume (210 mL) before cooling
to 50 °C and treating with TBME (210 mL). The mixture was
cooled to 10 °C and stirred for 1 h before isolating the product
3-Chloro-N-{(1S)-2-(1,3-dioxo-1,3-dihydro-2H-isoindol-2-
yl)-1-[(4-{8-[(1S)-1-hydroxyethyl]imidazo[1,2-a]pyridin-2-
yl}phenyl)methyl]ethyl}-4-[(1-methylethyl)oxy]benzamide (10):
Sodium hydrogen carbonate (211.4 g, 2.52 mol), 3-chloro-N-
{(1S)-2-[4-(chloroacetyl)phenyl]-1-[(1,3-dioxo-1,3-dihydro-2H-
isoindol-2-yl)methyl]ethyl}-4-[(1-methylethyl)oxy]benzamide
19 (0.961 kg, 1.74 mol), and tetrabutylammonium bromide
(95.14 g, 0.295 mol) were suspended in acetonitrile (10 L) and
stirred mechanically. A solution of (1S)-1-(2-amino-3-pyridi-
nyl)ethanol 7 (0.258 kg, 1.87 mol) in acetonitrile (1.53 L) was
added to the stirred mixture and the suspension heated to reflux.
The mixture was stirred at reflux for 26 h. The mixture was
cooled to 50 °C and treated with water (12.5 L) before cooling
to room temperature. The product was isolated by filtration,
washed with water (4.8 L), and dried under vacuum to yield
1
the title compound (1.02 kg, 92.6%) as a colorless solid: H
NMR (400 MHz, CDCl3) δH 1.36 (6H, d, J ) 6.0, (CH3)2),
1.72 (3H, d, J ) 7.0, CH3CH), 3.05 (1H, dd, J ) 14.0 and 7.0,
CH2Ph), 3.89 (2H, m, CH2NPhthal), 4.61 (1H, m, (CH3)2CHO),
4.71 (1H, m, CHNH), 5.27 (1H, q, J ) 7.0, CHOH), 5.39 (1H,
m, NH), 6.61 (1H, d, J ) 7.5, OH), 6.75 (1H, t, J ) 7.0, H-j),
6.91 (1H, d, J ) 7.5, H-b), 7.02 (1H, d, J ) 7.0, H-k), 7.34
(1H, d, J ) 8.0, H-d), 7.54 (1H, dd, J ) 7.5 and 2.0, H-c),
7.65-7.71 (2H, m, H-g), 7.75 (1H, s, H-a), 7.80-7.85 (3H,
m, H-h and H-f), 7.90 (2H, m, H-e), 8.03 (1H, d, J ) 8.0, H-i);
13C NMR (100 MHz, CDCl3) δC 21.9, 22.5, 38.6, 41.1, 50.8,
67.8, 71.9, 108.0, 112.4, 114.1, 119.7, 123.3, 123.8, 124.3,
126.3, 126.6, 127.1, 129.6, 129.8, 131.7, 132.0, 133.7, 134.1,
136.6, 144.2, 144.4, 156.2, 165.9 and 168.7; HRMS (ESI+) m/z
calcd for [M + H]+ C36H34N4O5Cl 637.2218, found 637.2238.
3-Chloro-N-{(1S)-2-[(N,N-dimethylglycyl)amino]-1-[(4-{8-
[(1S)-1-hydroxyethyl]imidazo[1,2-a]pyridin-2-yl}phenyl)-
methyl]ethyl}-4-[(1-methylethyl)oxy]benzamide(GSK923295A).
Deprotection Step: Hydrazine monohydrate (660 mL, 681.5 g,
13.61 mol) was added dropwise to a stirred suspension of
3-chloro-N-{(1S)-2-(1,3-dioxo-1,3-dihydro-2H-isoindol-2-yl)-
1-[(4-{8-[(1S)-1-hydroxyethyl]imidazo[1,2-a]pyridin-2-
yl}phenyl)methyl]ethyl}-4-[(1-methylethyl)oxy] benzamide 10
(867.4 g, 1.361 mol) in TBME (10.4 L) and IMS-74OP (3.34
L) at 40 °C over 20 min. The reaction mixture is stirred over
1262
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

by filtration and washing with 2-Me THF/TBME (1/1, 90 mL).
The cake was dried under vacuum to yield the title compound
(25.37 g, 90.9%) as a colorless solid that was spectroscopically
identical to that reported above.
Acknowledgment
We thank Dr. James Wickens and Mrs. Julia L. Beaurain
for analytical support throughout the project.
Supporting Information Available
SYNOPSIS TOC
Experimental procedures, characterization data and 1H NMR
are available for all isolated compounds. This information is
available free of charge Via the Internet at http://pubs.acs.org/.
The rapid discovery and scale-up of a new greener and robust
route to the CENP-E inhibitor GSK923295A originating from
the cheap and widely L-phenylalaninol is discussed. Also
discussed are the route derived impurities, how they were
unambiguously prepared to confirm structure and processing
amendments to control their formation, as well as enhancements
to the new process to facilitate future processing.
Received for review July 1, 2010.
OP100186C
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1263