Implementation of a high-temperature claisen approach for early phase delivery of a benzopyran derivative

Article abstract of DOI:10.1021/op900193v

Control of trace inorganics and an understanding of reaction kinetics and impurity generation proved critical in allowing safe and reliable scale-up of a high-temperature Claisen reaction in support of multikilogram manufacture of a key chromene intermediate 4. Implementation of a Medicinal Chemistry route allowed for rapid early phase delivery of 1 to support early clinical studies.

Full text of DOI:10.1021/op900193v

Organic Process Research & Development 2010, 14, 85–91
Implementation of a High-Temperature Claisen Approach for Early Phase Delivery of
a Benzopyran Derivative
Andrew J. Walker,*,‡ Sven Adolph,† Richard B. Connell,‡ Klaus Laue,† Michael Roeder,† Carsten J. Rueggeberg,§
Dirk Uwe Hahn,| Kurt Voegtli,⊥ and John Watson‡
Carbogen-AMCIS AG, Neulandweg 5, CH-5502, Hunzenschwil, Switzerland, GlaxoSmithKline Medicines Research Centre,
Gunnels Wood Road, SteVenage, Herfordshire SG1 2NY, United Kingdom, Carbogen-AMCIS AG, Hauptstrasse 171,
CH-4416, Bubendorf, Switzerland, and Carbogen-AMCIS AG, Schachenallee 29, CH-5001, Aarau, Switzerland
Abstract:
Chemistry route for large-scale synthesis to facilitate the
most expedient supply of initial quantities of active
pharmaceutical ingredient (API). In a subsequent article,
we will relate our work on identification of improved
alternative manufacturing routes to the benzopyran
core.2
Control of trace inorganics and an understanding of reaction
kinetics and impurity generation proved critical in allowing safe
and reliable scale-up of a high-temperature Claisen reaction in
support of multikilogram manufacture of a key chromene inter-
mediate 4. Implementation of a Medicinal Chemistry route allowed
for rapid early phase delivery of 1 to support early clinical studies.
Results and Discussion
A review of the Medicinal Chemistry route (Scheme
1) highlighted a number of key areas for further consid-
eration: thermal stability of propargyl ether 3 and chromene
4, general operability of the Claisen reaction, chromato-
graphic purification of 1 following capricious amide
formation, and numerous concentrations to dryness from
large [>100] volumes.
Introduction
5-Amino-6-bromo-N-[(1-(tetrahydro-2H-pyran-4-ylmethyl)-
4-piperidinyl]methyl-3,4-dihydro-2H-chromene-8-carboxam-
ide [1] is a 5-HT4 receptor agonist discovered at GlaxoSmith-
Kline1 and has been evaluated in preclinical studies. Agonists
for the 5-HT4 receptor have potential utility in the treatment of
a range of gastrointestinal disorders such as constipation, irritable
bowel syndrome, and functional dyspepsia.1
Use of chloro analogue 2 as starting material was driven
by (i) the potential requirement for preparation of a chloro
analogue of 1 at the outset of the project, (ii) literature
precedent that the presence of a halogen atom at this
position was beneficial to the thermal cyclisation,3 and
(iii) availability of this intermediate in large scale [>100
kg] from a related project. This approach gave us
maximum flexibility to meet the overall needs of the
program. The acetylene ether derivative of 3, in which
the acetyl group is replaced by pivaloyl, was reported to
give an improved profile for cyclisation,3 but this was not
evaluated due to lead time/availability of raw materials
and the potential need for an additional deprotection stage.
The propargyl ether 3 was prepared using a modifica-
tion of the Medicinal Chemistry process, wherein sodium
hydride and DMF/THF at reflux was replaced by potas-
sium carbonate in DMF at 70 °C. ARC data generated
for the full batch reaction [-19 J/g] revealed this to be a
safe and robust process for scale-up. Initial batches of
propargyl ether 3 were isolated following a simple water
drown-out and were found to contain significant levels of
inorganics [1-2% w/w]. This reduced the thermal stability
profile of 3, as judged by DSC, and resulted in significant
levels of impurities and very poor isolated yield [<20%
Initial quantities of 1 were prepared using the Medicinal
Chemistry route (Scheme 1) to support preliminary in Vitro
and in ViVo studies, and there was a need for rapid scale-
up of 1 to support further preclinical and early phase I
clinical studies. Assessment of the Medicinal Chemistry
route highlighted a number of key issues, and these were
rapidly assessed to see if this route was amenable to the
manufacture of multikilogram quantities of API. Herein,
we report our results in this direction, and describe how
we met the challenges of scaling up the Medicinal
* To whom correspondence should be addressed. E-mail: Andrew.J.Walker@
gsk.com.
† Carbogen-AMCIS AG, Hunzenschwil, Switzerland.
‡ GlaxoSmithKline Medicines Research Centre.
§ Carbogen-AMCIS AG, Bubendorf, Switzerland.
⊥ Carbogen-AMCIS AG, Aarau, Switzerland.
| Current address: Solvias, WRO-1060.P.14, Postfach, 4002 Basel, Switzerland.
(1) (a) Ahmed, M.; Miller, N. D.; Moss, S. F.; Sanger, G. J. WO
2007096352, 2007. (b) De Mayer, J.; Lefebvre, R.; Schuurkes, J.
Neurogastroenterol. Motil. 2008, 20 (2), 99. (c) For tegaserod, see:
Seerden, T. C.; De Man, J. G.; Holzer, P.; Van Den Bossche, R. M.;
Herman, A. G.; Pelckmans, P. A.; De Winter, B. Y. Neurogastroenterol.
Motil. 2007, 19 (10), 856.
(2) Rassias, G.; Stevenson, N. G.; Curtis, N. R.; Northall, J. M.; Gray, M.;
Prodger, J. C.; Walker, A. J. Org. Process Res. DeV. 2010, 14, DOI:
10.1021/op900188v.
(3) (a) Kakigami, T.; Baba, K.; Usui, T. Heterocycles 1998, 48 (12), 2611–
2619. (b) Baba, Y.; Usui, T.; Kakiue, T.; Baba, K. JP 08157466, 1996.
10.1021/op900193v  2010 American Chemical Society
Published on Web 12/07/2009
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
85

Scheme 1. Medicinal Chemistry route to 1
th] in the subsequent high-temperature cyclisation.4 Con-
sequently, an additional extractive water wash, followed
by solvent exchange, was incorporated for scale-up to give
3 of high purity and low residual inorganic content [<0.1%
w/w]. This was scaled up to 25 kg input to give 3 in 70%
th, >99.5% by HPLC.
Scheme 2. Impurities from Claisen rearrangement
Ether 3 was initially subjected to the existing thermal
cyclisation conditions [Dowtherm A, 220 °C] to assess the
reaction profile.
Use of alternative high-boiling solvents [NMP, DMSO,
dimethylacetamide, or ethylene glycol5] preferentially
generated different products [including 12 and 13] rather
than the desired product, and we reverted to Dowtherm
A to assess the potential of this system to provide a safe
and scaleable process. An isothermal kinetic study at
180-240 °C demonstrated that a minimum of >200 °C
was required to generate useful levels of the chromene 4
within a reasonable time frame for processing [<3 h]
without extensive concomitant decomposition of the
product.
Figure 1. HPLC plot of the high-temperature cyclisation of
alkyne ether 3 to give chromene 4.
A continuous chemistry approach to this stage was also
considered, as this had the potential to minimise the time during
which the reaction mixture was held at elevated temperature.
Unfortunately, levels of impurities rose in parallel with conver-
sion to 4, and dilution [required to provide a window for solution
flow chemistry] had a deleterious effect on the impurity profile.
The optimum conditions were established for a standard
batch process using Dowtherm A (5 vol) as solvent and heating
at 220-230 °C. Further profiling of the reaction generated
consistent results in lab experiments for both reaction conversion
and impurity profile, and provided confidence in the time
required to hold the reaction at 220-230 °C [45-60 min] prior
to workup. The observed decomposition of the reaction mixture
with extended heating required the batch to be cooled down
HPLC timecourse analysis revealed that the chromene
4 was unstable at the temperature required for cyclisation,
with a maximum of 67% (Figure 1, % a/a by HPLC)
observed after approximately 1 h. This decreased on
isothermal ageing to 39% after 4 h with an increase in
known impurities, benzofuran 12 and indole 13 (Scheme
2, as suggested by MS and literature3) alongside numerous
unknown low-level by-products. 3 was not completely
consumed until ∼2 h at 220 °C, but impurities 12 and 13
were much more difficult to purge from the product at
this, and downstream, stages. Consequently, we focused
our efforts on minimising levels of these impurities at the
expense of conversion in the initial campaigns.
(4) Ishii, H.; Ishikawa, T.; Takeda, S.; Ueki, S; Suzuki, M.; Harayama, T.
Chem. Pharm. Bull. 1990, 38 (6), 1775–1777 for an example of a base-
mediated alternative pathway.
(5) (a) Ariamala, G.; Balasubramanian, K. K. Tetrahedron 1989, 45 (1),
309–318. (b) Rao, U.; Balasubramanian, K. K. Tetrahedron Lett. 1983,
24 (45), 5023–5024.
86
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

rapidly after this time point. Given this narrow time-window,
it was recognised that on scale there would be insufficient time
to take an in-process check for HPLC conversion without
compromising the quality of the isolated product, and in-line
monitoring was not an option. After the in-process samples were
taken for retrospective analysis, the product was isolated by
addition of heptane to the Dowtherm mixture at 40 °C to
precipitate the product. Analysis demonstrated that yields of
50-60% could be obtained with an HPLC purity of 80% a/a.
Levels of unreacted 3 were typically 10-15% a/a in the isolated
product, and we also observed up to 10% mass balance
discrepancy between HPLC purity [80-90% a/a] and NMR
% w/w assay [70-80% w/w]. Downstream assessment showed
that we were able to proceed with this quality of intermediate
for initial supplies. This was demonstrated across a number of
lab batches to provide assurance in the proposed operating
conditions for the Claisen process. Recrystallisation of chromene
4 was evaluated, but the product could only be isolated in poor
overall yield and with modest improvement in purity [20-30%
th overall, 90-93% a/a] and was therefore not implemented
on scale.
Great care is always required when evaluating high
temperature scale-up reactions involving highly-energetic
groups such as acetylenes. Initial thermal screening by
DSC of ether 3 isolated from the Medicinal Chemistry
process had highlighted significant exothermic activity
from 140-300 °C [877 J/g, with sharp exotherm at 258
°C]. Implementation of the revised process for preparation
of 3 gave an improvement in its thermal stability as a
solid by DSC [exothermic activity from 181-299 °C, 453
J/g] due to the removal of inorganic impurities. Subsequent
ARC assessment of the final batch process [in 5 vol of
Dowtherm A] indicated we had a safe and robust system
for operation on scale using inorganic-free 3. An exotherm
of 95 J/g was observed at 180 °C, broadly consistent with
the onset for rearrangement, and no other thermal activity
was observed until 298 °C [exotherm of 35 J/g]. Two-
thirds of this initial energy was used to warm up the
reaction back up to 220 °C following addition of 3 to hot
Dowtherm A. The remaining low energy [∼30 J/g]
associated with the desired reaction in the ARC experiment
and the boiling barrier at 257 °C indicated that the process
was suitable for scale-up to 10 kg and beyond. However,
the sensitivity of the quality of the isolated product to
prolonged heating/cooling probably limits the utility of
this process for further scale-up.
propargyl ether 3 was converted to phenol 14 in this step rather
than the propyl ether analogue, and this was readily removed
in the following steps (Scheme 3). The purity of 5 could be
improved further by recrystallisation from tert-butyl methyl
ether, but since impurities were purged in the following steps,
it was decided to proceed directly with the bromination using
this quality of material.
Scheme 3. Tracking process impurities through the
synthesis
Modification to the bromination conditions involved addition
of a solution of 5 in dichloromethane to a suspension of NBS
in acetic acid. Impurity 14 was converted to the dibromophenol
15 which was effectively removed following isolation of 6 from
TBME (Scheme 3). On scale-up to plant, a delay in charging
5 to the NBS solution resulted in partial degradation of the
reagent, resulting in incomplete conversion. However, addition
of further NBS and workup gave 6 in good quality and assay
on 13-kg scale [96.5% a/a, 95.4% w/w assay, 63% th].
Importantly, we were now able to re-establish good analytical
agreement between HPLC and NMR assay of 6 at this point in
the synthesis. No benzylic bromination was observed during
this reaction.
Final saponification to cleave ester and amide protecting
groups was achieved in one pot using potassium hydroxide
in aqueous dioxane. Any remaining dibromo impurity 16
was efficiently purged during this process (Scheme 3),
and traces of any residual acetanilide acid were removed
following pH adjustment and isolation. The product 7 was
washed with ethyl acetate/heptane to provide control of
residual water in anticipation of the amide coupling [77%
th, 0.16% KF, 99.35% a/a by HPLC].
The diamine side chain 11 was prepared from 4-cyan-
opiperidine 9 and 4-formylpyran using a telescoped
reductive amination-nitrile reduction. Initial attempts to
perform this Via sequential hydrogenation were unsuc-
cessful, and we reverted to the use of LiAlH4 for reduction
of nitrile 10 to the primary amine 11. This was best
achieved using THF as solvent, and this solvent was
integrated into the reductive amination to provide a
streamlined process. Removal of aluminium salts and
isolation as its dihydrochloride salt from dioxane gave
11 in excellent yield and quality across three batches
[77-87% th, 96-98% w/w by NMR assay, 1.6-2.9%
w/w dioxane]. [Note - whilst the use of solvents such as
dichloromethane and dioxane are unsuitable for longer-
term manufacture due to their toxicity and environmental
impact, their utility as general process solvents for phase
I supply are considered acceptable.]
The batch process was successfully operated on mul-
tiple 10-kg scale to give the chromene 4 in the expected
yield and quality [53-59% th uncorr., 82-92% a/a,
75-84% w/w by NMR assay].
With the key fragment 4 in hand, we were able to assess
the remainder of the route to anilinoacid 7. Hydrogenation to
remove the double bond and effect dechlorination was readily
achieved using a modified EtOH/THF/Et3N system. These
solvents allowed for a much-simplified isolation of benzopyran
5 Via concentration and water antisolvent addition. This also
effectively removed residual triethylamine hydrochloride which
was a problem in the original process. Interestingly, residual
With quantities of the amine and acid in hand, we examined
the final amide coupling of 7 with 11. This was originally
achieved in Medicinal Chemistry using EDCI and catalytic
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
87

Scheme 4. N-acylurea
were identified at the outset of the project, notably safety and
purity concerns about the high-temperature Claisen reaction and
a capricious amide formation, as well as isolation and purifica-
tion procedures. This was supported by a good understanding
of purity using NMR assay as well as HPLC, and knowledge
of process impurities and their fate through the synthetic
sequence. This enabled rapid delivery of initial quantities of 1
in modest overall yield and excellent purity, and allowed us to
investigate alternative routes off the critical path for delivery.
These were focused on identification of improved routes to the
key benzopyran core and will be detailed in the subsequent
paper.2
DMAP, and the crude free base required extensive chromatog-
raphy to reduce levels of the related N-acylurea impurity
(Scheme 4).
This was typically formed in ∼10% during the reaction,
and it was also noted that conversion of 7 was also
incomplete. A number of standard peptide coupling agents
were evaluated [activation using SOCl2, cyanuric chloride,
CDI] before 2-chloro-4,6-dimethoxy-1,3,5-triazine (CD-
MT6) was selected on the grounds of optimum conversion
and operational simplicity. Initial solvent screening identi-
fied nitromethane and acetonitrile as preferred solvents,
and acetonitrile also facilitated the direct isolation of the
HCl salt of 1 from the reaction mixture by filtration,
removing the need for chromatography. A limited DoE
study [stoichiometry of amine (1-1.5 equiv), NEM (3-7
equiv), and CDMT (1-1.6 equiv), and temperature
(25-45 °C) in MeCN] indicated a maximum conversion
of 98% could be obtained at higher loadings of CDMT
and NEM, although in practice this did not translate to a
very high isolated yield in verification experiments. This
implied that there were additional factors that were not
considered in the initial design or in the workup, and will
be the subject of further work. Whilst conversion was not
complete, isolated yield and purity of the isolated HCl
salt was acceptable for onward processing, and this was
scaled-up to give the HCl salt, 8 in 76% th, 96.1% a/a on
a 9-kg scale.
The tosylate salt had previously been selected as the
version for development. Initially, this was prepared by
regeneration of the free base followed by salt formation
in acetonitrile. However, the lack of solubility of the
tosylate salt of 1 in acetonitrile precluded the development
of a controlled crystallisation, and we had additional
concerns regarding the use of acetonitrile in the final stage
[ICH class 2 solvent].7 Solvent screening identified DMSO
and isopropyl acetate as an appropriate binary system that
gave control of form and purity. A charcoal treatment to
reduce colour was also incorporated into the final process.
This was successfully scaled to 9 kg scale and provided
the desired salt of 1 in excellent overall yield and quality
to support phase I clinical supplies.
Experimental Section
General Remarks
The NMR spectra were recorded on 400 MHz Bruker
1
Advance for H NMR and 100 MHz for 13C NMR. HPLCs
were measured on Agilent Series 1100 (Agilent) or TSP
(Thermo) HPLC systems at 277 nm. Accurate MS spectra were
recorded on a Waters Micromass Q-ToF Ultima mass spec-
trometer. The ionisation technique was positive ion electrospray
with external lockmass correction. DSC analysis was performed
on the Mettler Toledo DSC 822e instrument.
Methyl 4-(Acetylamino)-5-chloro-2-(2-hydroxy)benzoate,
2 [CAS Registry no. 24190-77-0, prepared according to ref
1a]. NMR data for 2: 1H NMR (DMSO-d6, 400 MHz, 300 K):
10.56-10.47 (s, 1H, NH), 9.55-9.45 (s, 1H, 2-OH), 7.80-7.77
(s, 1H, 6-H), 7.74-7.71 (s, 1H, 3-H), 3.86-3.81 (s, 3H,
H3COC(O)), 2.22-2.12 (s, 3H, H3CC(O)NH); 13C NMR
(DMSO-d6, 400 MHz, 300 K): 167.7, 159.3 (C-2), 141.0 (C-
4), 130.5 (C-6), 114.2 (C-5), 111.5 (C-1), 110.1 (C-3), 52.8
(H3COC(O)), 24.4 (H3CC(O)NH), 93-95% a/a by HPLC.
Methyl 4-(Acetylamino)-5-chloro-2-(2-propynyloxy)ben-
zoate, 3. A suspension of methyl 4-(acetylamino)-5-chloro-2-
hydroxybenzoate 2 (24.78 kg, 101.7 mol) and potassium
carbonate (16.85 kg, 121.9 mol) in DMF (247 L) was treated
with propargylbromide (80% in toluene, 15.87 kg, 106.7 mol)
at rt. The resulting reaction mixture was stirred at rt before
warming to 70 °C and ageing for 5.5 h (98% conversion by
HPLC). The reaction mixture was cooled to 35 °C and treated
with water (295 L). The resulting suspension was cooled to rt
and stirred overnight. The product was filtered off, and the filter
cake was rinsed three times with water (3× 25 L).
The resulting wet cake was dissolved in dichlo-
romethane (245 L) and methanol (12 L) and washed with
water (2× 42 L). The organic solution was concentrated
to ∼75 L, and a solvent switch from dichloromethane to
TBME was performed using 3× 51 L. The resulting
suspension was stirred at rt and filtered off. The filter cake
was twice rinsed with TBME (21 L) and dried to give 3
as a pale-brown solid (19.73 kg, 99.8% a/a, <0.1% ROI),
In summary, scaleable processes using the existing Medicinal
Chemistry route were developed for the synthesis of 1 on
multikilogram scale. This addressed a number of issues that
1
DSC mp: 145.2 °C. H NMR (CDCl3, 400 MHz, 300 K):
(6) (a) Kaminski, Z. J. Synthesis 1987, 917–920. (b) Garrett, C. E.; Jiang,
X.; Prasad, K.; Repic, O. Tetrahedron Lett. 2002, 43, 4161–4165. (c)
Kaminski, Z. J.; Paneth, P.; Rudzinski, J. J. Org. Chem. 1998, 63, 4248–
4255. (d) Review: Blotny, G. Tetrahedron 2006, 62, 9507–9522.
(7) ICH Harmonised Tripartite Guideline: Impurities: Guideline for
Residual SolVents Q3C, (R3), Step 4; International Conference on
Harmonisation, 2005. http://www.ich.org/LOB/media/MEDIA423.pdf.
δ 8.54-8.41 (s, 1H, NH), 7.96-7.86 (s, 1H, 6-H),
7.86-7.64 (br s, 1H, 3-H), 4.89-4.77 (d, J ) 4 Hz, 2H,
OCH2CCH), 3.95-3.82 (s, 3H, H3COC(O)), 2.62-2.51
(t, J ) 4 Hz, 1H, OCH2CCH), δ ) 2.36-2.22 (s, 3H,
H3CC(O)NH); 13C NMR (CDCl3, 400 MHz, 300 K): δ
88
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

168.7, 164.7, 156.9 (C-2), 138.8, 132.0, 115.9, 113.6 (C-
1), 106.2 (C-3), 77.5 (CH2CCH), 76.4 (CH2CCH), 57.1
(H3COC(O)), 52.2 (OCH2CCH), 25.2 (H3CC(O)NH). MS
calculated for C13H12ClNO4 282.0533 (MH+), found
282.0534.
was stirred overnight at 20 °C, and HPLC analysis
revealed only 71% a/a conversion. Further N-bromosuc-
cinimide was added (2.34 kg, 13.1 mol), and the mixture
was further stirred for 2 h, giving 90% a/a conversion
according to HPLC. Water (66 L) was added, and the
resulting organic phase was washed further with dichlo-
romethane (66 L) and water (66 L). The organic phase
was concentrated under reduced pressure at 40 °C to ∼66
L, and TBME (132 L) was added to give an off-white
suspension. This was cooled further to 0-5 °C, filtered,
washed with TBME (2× 40 L), and dried under nitrogen
to give 6 as an off-white solid (10.93 kg, 63% th, 96.1%
Methyl 5-(Acetylamino)-6-chloro-2H-chromene-8-car-
boxylate, 4. Dowtherm A (36.8 kg diphenylether and 15.8
kg biphenyl) was heated to 220 °C and 3 (9.61 kg, 35.5
mol) was added Via a hastelloy drum in a single solid
addition, causing a temperature drop down to 195 °C. The
mixture was heated up to 220 °C over 11 min, and the
solution was maintained at 220 °C for 50 min before it
was cooled down to 40 °C. Heptane (96 L) was added at
40 °C, and the resulting suspension was cooled to 20 °C
and stirred overnight at this temperature. The product was
filtered, washed with heptane (2× 6.8 L) and MIBK (2×
6.8 L), and dried under a nitrogen stream gave 4 as a
brown solid (5.44 kg, 92.3% a/a, 84.3% w/w NMR assay).
1H NMR (CDCl3, 400 MHz, 300 K): δ 7.75-7.71 (s, 1H,
NH), 7.15-7.03 (br s, 1H, 7-H), 6.38-6.28 (d, J ) 12
Hz, 1H, 4-H), 5.97-5.86 (m, 1H, 3-H), 4.95-4.88 (t, J
) 4 Hz, 2H, 2-H), 3.89-3.87 (s, 3H, H3COC(O)),
2.30-2.25 (s, 3H, H3CC(O)NH). DSC melt: Peak 189.2
°C. MS calculated for C13H12ClNO4 282.0533 (MH+),
found 282.0522.
1
a/a, 95.4% w/w NMR assay). H NMR (DMSO-d6, 400
MHz, 300 K): δ 9.72-9.63 (s, 1H, NH), 7.75-7.66 (s,
1H, 8-H), 4.23-4.04 (t, J ) 4 Hz, 2H, OCH2CH2CH2),
3.80-3.75 (s, 3H, H3COC(O)), 2.69-2.54 (m, 2H,
OCH2CH2CH2), δ 2.11-1.98 (s, 3H, H3CC(O)NH), δ
1.94-1.79 (t, J ) 4 Hz, 2H, OCH2CH2CH2); 13C NMR
(DMSO-d6, 400 MHz, 300 K): δ 168.2, 165.1, 153.9 (a-
C), 139.4 (C-5), 131.1 (C-7), 124.8 (C-6), 119.7 (b-C),
112.1 (C-8), 66.4 (OCH2CH2CH2), 52.5 (H3COC(O)), 22.9
(H3CC(O)NH), 22.3 (OCH2CH2CH2), 20.8 (OCH2CH2-
CH2). DSC melt: Peak 199.7 °C. MS calculated for
C13H14BrNO4 328.0184 (MH+), found 328.0191.
5-Amino-6-bromo-3,4-dihydro-2H-chromene-8-carbox-
ylic Acid, 7. A suspension of 6 (10.86 kg, 33.1 mol) in
dioxane (55 L) and water (55 L) was treated with a
solution of potassium hydroxide (15.3 kg, 272.7 mol) in
water (117 L) at ambient temperature, and the resulting
suspension was heated to 90 °C overnight. The solution
was cooled to 20 °C and treated with 5 N HCl (47 L) to
pH 5.6 to give a white suspension. This was cooled,
filtered, and washed with water (2× 55 L) and then ethyl
acetate/heptane (4× 5 vol, 1:3 mixture). The cake was
dried in a conical dryer at 50 °C under vacuum to give 7
as an off-white solid (6.93 kg, 77% th, 99.4% a/a, 99.4%
Methyl 5-(Acetylamino)-3,4-dihydro-2H-chromene-8-car-
boxylate, 5. A suspension of 4 (22.31 kg, 79.4 mol) and
10% Pd/C (Degussa E101 R/W, 50% wet, 4.45 kg) in
absolute ethanol (223 L) and THF (223 L) was dosed with
triethylamine (8.42 kg, 83.7 mol), the mixture was purged
with hydrogen (e1200 mbar) and stirred for 18 h at 20
°C. The reaction mixture was filtered through prewashed
Celite (22.3 kg) and washed with ethanol/THF (46 L, 1:1).
The filtrate was concentrated to ∼70 L vol at 45-50 °C
under reduced pressure, cooled to 20 °C, and treated with
water (225 L). The resulting suspension was stirred at this
temperature for 2 h before cooling further to 5-10 °C.
The suspension was filtered, washed with water (45 L)
and ethyl acetate/heptane (1:3 ratio, 2× 46 L), and dried
to give 5 as a pale-brown solid (13.22 kg, 67% th, 81.9%
1
w/w NMR assay, KF 0.03% w/w), H NMR (DMSO-d6,
400 MHz, 300 K): δ 11.88-11.76 (s, 1H, COOH),
7.70-7.60 (s, 1H, 8-H), 5.78-5.60 (br s, 2H, NH2),
4.18-4.00 (t, J ) 4 Hz, 2H, OCH2CH2CH2), 2.48-2.40
(m, 2H, OCH2CH2CH2), 2.01-1.84 (t, J ) 4 Hz, 2H,
OCH2CH2CH2); 13C NMR (DMSO-d6, 400 MHz, 300 K):
δ 165.9 (COOH), 155.7 (a-C), 148.0 (C-5), 133.3 (C-7),
108.5 (C-6), 107.9 (C-8), 97.8 (b-C), 65.8 (OCH2CH2CH2),
21.5 (OCH2CH2CH2), 21.0 (OCH2CH2CH2). DSC melt:
Peak 220.3 °C. MS calculated for C10H10BrNO3 271.9922
(MH+), found 271.9918.
1-Tetrahydro-2H-pyran-4-ylmethyl-4-piperidinemetha-
namine Dihydrochloride, 11. A solution of 4-cyanopiperi-
dine 9 (5.08 kg, 46.1 mol) and 4-formyltetrahydropyran
(5.26 kg, 46.1 mol) in THF (102 L) was treated with acetic
acid (1.0 L) at 22 °C and aged for 30 min. Ten percent
Pd/C (1.02 kg) was added, the reactor was purged with
hydrogen, and the mixture was hydrogenated at 20 °C
overnight. Anhydrous Na2SO4 (5.09 kg) was added to the
purged reaction mixture; the mixture was stirred for 30
min and then filtered off using prewashed Celite (5.09
kg), washing through with further THF (10 L). The filtrate
1
a/a, 84.5% w/w NMR assay). H NMR (DMSO-d6, 400
MHz, 300 K): δ 9.36-9.25 (s, 1H, NH), 7.48-7.43 (d, J
) 12 Hz, 1H, 8-H), 7.22-7.12 (d, J ) 8 Hz, 1H, 7-H),
4.18-4.10 (t, J ) 4 Hz, 2H, OCH2CH2CH2), 3.78-3.70
(s, 1H, H3COC(O)), δ 2.68-2.58 (m, 2H, OCH2CH2CH2),
δ 2.12-2.07 (s, 3H, H3CC(O)NH), δ 1.94-1.86 (t, J ) 4
Hz, 2H, OCH2CH2CH2); 13C NMR (DMSO-d6, 400 MHz,
300 K): δ 168.9, 166.3, 154.9 (a-C), 141.0 (C-5), 128.6
(C-7), 116.9 (C-8), 115.6 (C-6), 115.3 (b-C), 66.1
(OCH2CH2CH2), 23.9 (OCH2CH2CH2), 21.2 (OCH2CH2-
CH2), 20.9 (H3CC(O)NH). DSC melt: Peak 171.8 °C. MS
calculated for C13H15NO4 250.1079 (MH+), found 250.1069.
Methyl
5-(Acetylamino)-6-bromo-3,4-dihydro-2H-
chromene-8-carboxylate, 6. A solution of 5 (13.21 kg, 53.1
mol) in dichoromethane (132 L) was prepared in the feed
tank and added over 3 h to a fine pale-yellow suspension
of N-bromosuccinimide (10.40 kg, 58.4 mol) in acetic acid
(66 L), prepared at 20 °C. The clear orange-red solution
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
89

was cooled to 0-5 °C and dosed with 2.4 M lithium
aluminium hydride in THF solution (26.2 kg) over ∼2 h
and warmed to ambient overnight. Analysis indicated
incomplete conversion, and thus further 2.4 M lithium
aluminium hydride in THF solution (8.76 kg) was added
at 0-5 °C. The mixture was warmed to ambient temper-
ature and stirred for a further 1 h, giving full conversion
by GC. The reaction mixture was cooled down to -5 °C,
and water (9.6 L) was added carefully. The quenched
reaction mixture was warmed to ambient, and 30% NaOH
solution (1.5 L) was added. Anhydrous Na2SO4 (10.17 kg)
was charged to the purged reactor and was stirred for 30
min, filtered, and washed with THF (2× 31 L). A solvent
switch from THF to dioxane was performed at 40 °C under
reduced pressure, and the resulting suspension was cooled
to 10 °C and treated with 4 M HCl in dioxane (35 L) to
give the salt as a white suspension. The white suspension
was stirred and warmed to ambient temperature and was
filtered and washed with dioxane (2× 20 L). The product
was dried on a rotary evaporator at 40 °C to give 11 as a
white solid (11.51 kg, 87% th, 98.4% w/w NMR assay,
residual dioxane 2.8% w/w, <0.1% ROI). 1H NMR (D2O,
400 MHz, 300 K): δ 3.89-3.81 (dd, J ) 4 Hz, 2H, 2-Ha,
6-Ha), 3.59-3.47 (dd, J ) 4 Hz, 2H, 2-Hb, 6-Hb),
3.40-3.31 (t, J ) 8 Hz, 2H, tetrahydropyran-CH2-bridge),
3.31-3.00 (2× dd, J ) 4 Hz, 1H, CHCH2CH2O),
2.94-2.82 (m, 5H, 8-Ha, 12-Ha, 8-Hb, 12-Hb, CHCH2NH2),
2.15-2.00 (m, 2H, 9-Hb, 12-Hb), 1.99-1.84 (m, 2H,
piperidine-CH2-bridge), 1.62-1.56 (d, J ) 12 Hz, 3-Hb,
5-Hb), 1.52-1.34 (d, J ) 12 Hz, 9-Ha, 11-Ha), 1.30-1.17
(m, 2H, 3-Ha, 5-Ha); 13C NMR (D2O, 400 MHz, 300 K):
δ 67.1 (2× CHCH2CH2O), 62.5 (piperidine-CH2-bridge),
52.7 (2× CHCH2CH2N), 49.6 (H2NCH2-), 43.7 (CHCH2-
CH2O), 31.8 (CHCH2CH2N), 29.8 (1× CHCH2CH2O),
29.7 (1× CHCH2CH2O), 26.6 (2× CHCH2CH2N). DSC
melt: Broad peak 90 °C. MS calculated for C12H24N2O
213.1967 (MH+ for parent), found 213.1958.
5-Amino-6-bromo-N-[(1-(tetrahydro-2H-pyran-4-ylmethyl)-
4-piperidinyl]methyl-3,4-dihydro-2H-chromene-8-carboxa-
mide, Monohydrochloride, 8. A suspension of 2H-chromene-
8-carboxylic acid 7 (6.85 kg, 25.17 mol), diamine · 2HCl
11 (9.09 kg, 31.87 mol), and 2-chloro-4,6-dimethoxy-
1,3,5-triazine (ex AlzChem, >98% purity, 7.03 kg, 40.04
mol) in acetonitrile (130.3 kg) was treated with N-
ethylmorpholine (20.2 kg, 175.2 mol) over 23 min at 23
°C, and the mixture was allowed to stir for 16 h at 22-25
°C. The mixture was filtered, washed with acetonitrile (3×
27.6 kg), and dried in Vacuo at 40 °C to give 8 as a beige
solid (8.93 kg, 76% th, 96.1% a/a HPLC). 1H NMR
(DMSO-d6, 400 MHz): δ 10.23 (broad s, 1H, R3NH+),
8.04 (t, J ) 4 Hz, 1H, amide NH), 7.73 (s, 1H, H8), 5.52
(s, 2H, aryl NH2), 4.21 (t, J ) 5 Hz, 2H, H2), 3.83 (dd,
J ) 12 Hz, J ) 2 Hz, 2H, H22, H24), 2.97-3.51 (m, 6H,
H15, H17, H19, H22, H24), 2.76-2.94 (m, 4H, H12, H15,
H17), 2.49 (t, J ) 6 Hz, 2 H, H4), 2.01-2.14 (m, 1 H,
H13), 1.96 (dt, J ) 5 Hz, 2H, H3), 1.57-1.85 (m, J ) 10
Hz, 7H, H14, H18, H20, H21, H25), 1.16-1.29 (m, 2H,
H14, H18). 13C NMR (DMSO-d6, 100 MHz): δ 164.0
(C11), 153.1 (C10), 146.4 (C6), 131.7 (C8), 111.2 (C5),
107.5 (C9), 98.5 (C7), 66.3 (C22, C24), 65.9 (C2), 61.6
(C12), 52.2 (C15, C17), 44.5 (C19), 34.4 (C20), 30.8 (C14,
C18), 29.8 (C13), 26.5 (C21, C25), 21.0 (C4), 20.7 (C3).
DSC melt: Peak 222.5 °C. MS calculated for
C22H32BrN3O3 466.1705 (MH+ for parent), found 466.1701.
5-Amino-6-bromo-N-[(1-(tetrahydro-2H-pyran-4-ylmethyl)-
4-piperidinyl]methyl-3,4-dihydro-2H-chromene-8-carboxa-
mide, Monotosylate, 1. The HCl salt 8 (8.89 kg, 17.68 mol)
was stirred for 2 h at ambient temperature in deionised water
(51 L), and the brown turbid solution was filtered. The filtrate
was cooled to 3 °C, and an aqueous sodium carbonate solution
(13.73% w/w; 20 L) was added over 70 min to adjust the pH
to 10. The resulting suspension was transferred in two portions
to the filter, and the cake was rinsed with water (3× 39.5 L).
The resulting free base was dried at 53 °C to give a residual
water content of 9.6% w/w by KF. The free base was dissolved
out of the filter using acetonitrile (30.3 kg), and the solution
was treated with activated carbon (0.17 kg) whilst warming to
69 °C. The charcoal was removed by filtration and rinsed with
hot acetonitrile (5 kg). A solution of 4-toluene sulfonic acid
monohydrate (3.15 kg, 16.56 mol) in acetonitrile (22.7 kg) was
added to the filtrate at 68-73 °C over 60 min. The reaction
mixture was cooled to 10 °C over 2 h and filtered, washing the
cake with acetonitrile (26.2 kg). The crude tosylate salt was
dried at 50 °C in Vacuo and then dissolved on the filter in
dimethyl sulfoxide (60.4 kg) directly on the nutsche at 90 °C.
The solution was filtered, and the filter was rinsed with hot
dimethyl sulfoxide (6 kg). DMSO filtrates were combined and
heated to 80 °C. Isopropyl acetate (71.5 kg) was added at 78-80
°C, and the clear solution was seeded with a suspension of 1
(20 g) in isopropyl acetate (0.3 kg). The turbid solution was
aged for 23 min, and further isopropyl acetate was added (51.2
kg) at 79-81 °C. The resulting suspension was cooled to 20
°C, aged, and then filtered. The cake was rinsed with isopropyl
acetate (2× 13.7 kg) and dried at 50 °C in Vacuo to give 1 as
a slightly coloured crystalline solid in (8.91 kg, 79% th, 99.8%
1
a/a HPLC, 0.29% w/w DMSO, 0.03% w/w IPAc). H NMR
(DMSO-d6, 400 MHz): δ 8.69 (broad s, 1H, R3NH+), 8.03 (t,
J ) 6 Hz, 1H, amide NH), 7.74 (s, 1H, H8), 7.49 (d, 2H, H29,
H31), 7.12 (d, J ) 8 Hz, 2H, H23, H32), 5.52 (broad s, 2H,
aryl NH2), 4.20 (t, J ) 5 Hz, 2H, H2), 3.79-3.87 (m, 2H, H22,
H24), 3.49 (d, J ) 12 Hz, 2H, H15, H17), 3.15-3.35 (m, 4H,
H19, H22, H24), 2.80-2.96 (m, J ) 6 Hz, 4H, H12, H15, H17),
2.48 (t, J ) 6 Hz, 2H, H4), 2.29 (s, 3H, H26), 1.99-2.09 (m,
1H, H13), 1.96 (dt, J ) 6 Hz, 2H, H3), 1.72-1.84 (m, 3H,
H20, H21, H25), 1.62 (d, J ) 11 Hz, 2H), 1.40-1.54 (m, 2H,
H21, H25), 1.14-1.27 (m, 2 H, H14, H18). 13C NMR (DMSO-
d6, 100 MHz): δ 164.1 (C11), 153.1 (C10), 146.4 (C6), 145.6
(C27), 137.7 (C30), 131.8 (C8), 128.0 (C23, C32), 125.5 (C29,
90
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

C31), 111.1 (C5), 107.5 (C7), 98.5 (C9), 66.2 (C22, C24), 65.9
(C2), 61.4 (C12), 52.3 (C15, C17), 43.6 (C19), 33.7 (C20), 30.2
(C14, C18), 29.6 (C13), 26.7 (C21, C25), 21.0 (C4), 20.7 (C3,
C26). DSC melt: Peak 229.4 °C.
Acknowledgment
We thank Fabian Hunziker, Peter Lazar, and Marco Roth
(Carbogen AMCIS) for their technical assistance.
Supporting Information Available
DOE results for CDMT amide coupling screen. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Received for review July 27, 2009.
OP900193V
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
91