Investigation of Synthetic Routes to a Key Benzopyran Intermediate of a 5HT4 Agonist

Article abstract of DOI:10.1021/op900188v

The supply route to GlaxoSmithKline's 5HT₄ receptor agonist 1 centred on the construction of key benzopyran fragment 2. Our attempts to define the final manufacturing route for this component are described through a series of disconnections. The systematic approach undertaken towards the construction of the benzopyran skeleton focused on cycli- sation strategies from appropriate precursors and evaluation of the performance of the key steps.

Full text of DOI:10.1021/op900188v

Organic Process Research & Development 2010, 14, 92–98
Investigation of Synthetic Routes to a Key Benzopyran Intermediate of a 5HT4
Agonist
Geracimos Rassias,* Neil G. Stevenson, Neil R. Curtis, John M. Northall, Matthew Gray, Jeremy C. Prodger, and
Andrew J. Walker
GlaxoSmithKline R&D, Chemical DeVelopment, Synthetic Chemistry, Gunnels Wood Road, SteVenage,
Hertfordshire SG1 2NY, U.K.
Abstract:
The supply route to GlaxoSmithKline’s 5HT4 receptor agonist
1 centred on the construction of key benzopyran fragment
2. Our attempts to define the final manufacturing route for
this component are described through a series of discon-
nections. The systematic approach undertaken towards the
construction of the benzopyran skeleton focused on cycli-
sation strategies from appropriate precursors and evalua-
tion of the performance of the key steps.
Figure 2. Disconnections applied to the key benzofuran
intermediate.
construct the benzopyran skeleton which warranted experimen-
tal investigation (Figure 2).
1. Introduction
Disconnection A requires a 2,3-dihydropyran to participate
as a dienophile in a cycloaddition reaction to efficiently form
the aromatic ring in 3 (Figure 2). In contrast, disconnections B
and C use an appropriately substituted benzene ring on which
the pyran ring is constructed. Disconnections B and C differ in
their cyclisation strategy, proceeding Via formation of either a
final carbon-carbon or a carbon-oxygen bond to construct the
benzopyran. The results from our efforts on disconnections A,
B, and C are described below.
GlaxoSmithKline’s 5-HT4 receptor agonist 1, 5-amino-6-
bromo-chroman-8-carboxylic acid [1-(tetrahydro-pyran-4-ylm-
ethyl)-piperidin-4-ylmethyl]-amide (Figure 1), has been evalu-
ated in several studies. Agonists for the 5-HT4 receptor have
potential utility in the treatment of gastrointestinal disorders such
as constipation, irritable bowel syndrome, and functional
dyspepsia.1
2. Results and Discussion
2.1. Disconnection A. The benzopyran skeleton could be
constructed by the Diels-Alder reaction between an ap-
propriately substituted dihydropyran (4a or 4b) and 3,6-
dichloropyridazine (5) or pyrones 6a and 6b (Scheme 1).3,4 The
Diels-Alder reaction with 3,6-dichloropyridazine would pro-
vide, after extrusion of nitrogen gas and aromatisation (e.g., by
loss of HBr or TMSOH), a dichlorobenzopyran intermediate,
while the reaction with the pyrone derivatives would allow
regiospecific introduction of both the amino and carboxylate
groups in 3. Initial scoping studies using 4a/b and 5 or 6a/b
suggested that the high-temperature conditions required for
Diels-Alder reaction were detrimental for dihydropyran sub-
strates 4a/b as well as for pyrones 6a/b (loss of the Boc group).
Figure 1. Target 5HT4 agonist and key benzopyran intermediate.
5-Amino-6-bromo-chroman-8-carboxylic acid, 2 (Figure 1),
is the key component of 1, and our early supply route was based
on a high-temperature Claisen rearrangement.2 In spite of the
success of this route in delivering early supplies, there were
significant operability and quality issues associated with the
Claisen rearrangement that made this route unsuitable for larger-
scale manufacturing. Thus, a detailed study on alternative routes
was initiated and targeted generic structure 3 as a suitable
precursor to 2. We considered three key disconnections to
Scheme 1. Attempted Diels-Alder reactions
* To whom correspondence should be addressed. E-mail:
geracimos.9.rassias@gsk.com.
(1) De Mayer, J.; Lefebvre, R.; Schuurkes, J. Neurogastroentero. Motil.
2008, 20 (2), 99. 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. Neurogastroentero. Motil. 2007, 19 (10), 856.
(2) Walker, A. J.; Adolph, S.; Connell, R. B.; Laue, K.; Roeder, M.;
Carsten, J.; Rueggeberg, C. J.; Hahn, D. U.; Voegtli, K.; Watson, J.
Org. Process Res. DeV. 2009. DOI: 10.1021/op900193v.
92
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development
10.1021/op900188v CCC: $40.75  2010 American Chemical Society
Published on Web 11/06/2009

Scheme 3. Preparation of aryl-propargyl ether substratesa
Attempts to promote the cycloaddition at lower temperatures
by the use of Lewis acids were also unsuccessful. In contrast,
dihydropyran 4c as stable under all reaction conditions attempted
but failed to give the desired Diels-Alder adducts.
2.2. Disconnection B. In disconnection B formation of the
benzopyran skeleton involves an etherification reaction as the
final step. The Heck reaction of 75 with allyl alcohol occurred
selectively at the iodo-position, but the resulting aldehyde 8
did not cyclise with the phenol terminus to form the desired
cyclic acetal (or dehydrate to the dihydropyran) (Scheme 2).
Instead, a 1:3 equilibrating mixture of 8 and 9 was isolated in
38% yield after chromatography. We reasoned that disconnec-
tion approach B (Figure 2) would be undermined by chemose-
lectivity issues since both the anilide and phenolic groups may
compete for the pendant activated intermediate. The option of
doubly protecting the aniline was considered, but the length of
the synthetic route drove us to prioritise more attractive
retrosynthetic options under disconnection C.
Scheme 2. Chemoselectivity issues in the cyclisation stepa
a Reagents and conditions: a) Pd(OAc)2, allyl alcohol, n-Bu4NCl, NMP,
55 °C.
a Reagents and conditions: a) Methanol, H2SO4, 90 °C, 3 h; b) RCOCl,
NaHCO3 (aq), EtOAc, rt; c) propargyl bromide, K2CO3, NMP at 90 °C or acetone
reflux; d) SO2Cl2, EtOAc, rt.
2.3. Disconnection C. 2.3.1. Metal-Catalysed Cycloisom-
erisation of Aryl-Propargyl Ethers. The strategy in disconnec-
tion C focused on accessing appropriately substituted aryl ether
intermediates that allowed regiospecific formation of the pyran
ring by reaction at the adjacent aromatic carbon atom.6 In recent
years there have been several literature reports regarding the
metal-catalysed cycloisomerisation of aryl-propargyl ethers to
chromenes as an alternative to the high-temperature Claisen
rearrangement.7 We decided to test this approach and investigate
the performance of various substrates and their interaction with
potential catalysts. In addition to our chloro-substrate 13c
(utilised in the Claisen route),2 we prepared the des-chloro
phenols 12a and 12b on multikilogram scale and converted
them to their corresponding propargyl ethers 13a and 13b
(Scheme 3).
Scheme 4. Metal-catalysed cycloisomerisation of
aryl-propargyl ethersa
a Reagents and conditions: see Table 1.
substrate 13c performed very poorly, presumably due to
solubility issues although stereoelectonic factors cannot be
discounted. Our best results were obtained in nonpolar solvents
such as 1,4-dioxane, 2-MeTHF, or toluene at 80-100 °C with
substrates 13a,b. Triphenylphosphine-gold(I) catalysis proved
superior to both platinum(II) and -(IV) reactions in terms of
yield and selectivity for the desired cycloisomerisation process.9
The main side reaction, which is more pronounced when the
platinum catalysts are employed, is the depropargylation of the
ether substrate, furnishing the parent phenol 12a-c as the main
impurity (Scheme 4). With substrates 13a,b the depropargyla-
tion reaction is significantly suppressed under the gold(I)
conditions. Reports exist in the literature where platinum(II)
chloride is optimal for this type of process although no rationale
has been offered to date.7,8 We reasoned that coordination of
Following the work of the Sames, Echavaren, and Gagosz
groups,7,8 we screened our substrates against platinum(II) and
-(IV) chlorides as well as gold(I) complexes (Table 1). Chloro-
(3) Barluenga, J.; Vazquez-Villa, H.; Merino, I.; Ballesteros, A.; Gonzalez,
J. M. Chem.sEur. J. 2006, 12 (22), 5790.
(4) (a) Broch, S.; Anizon, F.; Moreau, P. Synthesis 2008, 13, 2039. (b)
Mouaddib, A.; Joseph, B.; Hasnaoui, A.; Merour, J.-Y. Synthesis 2000,
4, 549.
(5) Intermediate 7 was prepared in one step from the des-iodo analogue
which was used in the first supply route of 1; see reference 2.
(6) Parham, W. E.; Bradsher, C. K. Acc. Chem. Res. 1982, 15, 300.
(7) (a) Pastine, S. J.; Youn, S. W.; Sames, D. Org. Lett. 2003, 5, 1055.
(b) Pastine, S. J.; Youn, S. W.; Sames, D. Tetrahedron 2003, 59, 8859.
(c) Martin-Mahute, B.; Nevado, C.; Ca´rdenas, D. J.; Echavarren, A. M.
J. Am. Chem. Soc. 2003, 125, 5757.
(8) (a) Nevado, C.; Echavarren, A. M. Chem.sEur. J. 2005, 11, 3155.
(b) Me´zailles, N.; Ricard, L.; Gagosz, F. Org. Lett. 2005, 7, 4133.
(9) Curtis, N. R.; Prodger, J. C.; Rassias, G.; Walker, A. J. Tetrahedron
Lett. 2008, 49, 6279.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
93

Table 1. Results from substrate and catalyst screening in the metal-catalysed cycloisomerisation of aryl propargyl ethers
HPLC ratio (% a/a)
substrate
catalyst (mol %)
conditions
14
12
15
13a
13a
13a
13a
13aa
13a
13b
13b
13b
13b
13b
13cb
PtCl4 (5)
PtCl2 (5)
PtCl2 (5)
PtCl2 (5)/AgOTf (5)
AgOTf (5)
(Ph3P)AuNTf2 (0.5)
(Ph3P)AuNTf2 (0.5)
(Ph3P)AuNTf2 (0.5)
(Ph3P)AuNTf2 (0.5)
(Ph3P)AuNTf2 (0.1)
(Ph3P)AuCl/AgOTf (1)
(Ph3P)AuCl/AgOTf (2)
1,4-dioxane, 55 °C, 18 h
1,4-dioxane, 85 °C, 18 h
2-MeTHF, 70 °C, 24 h
toluene, 100 °C, 24 h
2-MeTHF, 75 °C, 18 h
toluene, 85 °C, 1 h
2-MeTHF, 70 °C, 5 h
CF3C6H5, 85 °C, 1 h
toluene, 85 °C, 1 h
27.8
50.2
53.4
50.5
47.0
92.1
84.1
90.5
91.9
94.0
79.1
10.7
49.1
25.9
23.5
21.8
8.0
3.1
7.9
2.8
2.5
4.1
2.0
1.8
2.0
2.7
1.3
3.7
2.2
1.5
0.9
6.0
2.1
toluene, 85 °C, 2 h
2-MeTHF, 70 °C, 3 h
2-MeTHF, 70 °C, 3 h
2.3
11.6
5.5
a Mass balance is 16 and 17. b Mass balance is unreacted starting material.
the phenolic oxygen atom to the catalyst (assisted by the
neighbouring ester group) could compete with the desired
coordination of the alkyne moiety to the catalyst. This metal
complexation of the ethereal oxygen atom could activate the
methylene group of the propargyl ether towards nucleophilic
attack by the halide anions present in the metal salt. For this
reason we decided to include silver(I) triflate in an attempt to
sequester the chloride anions. This modification accelerated the
rate of the desired reaction, but the extent of depropargylation
was not suppressed. On the basis of this result we reasoned
that the depropargylation reaction is probably not chloride
assisted. Instead we propose that the propargyl ether isomerises
to the allenyl enol ether,10 which in turn hydrolyses into the
parent phenol and acrolein under the hplc conditions we use to
monitor the reaction.
Interestingly, silver(I) triflate alone catalysed the cycloisom-
erisation of our substrates, but the product distribution was
typical of the high-temperature Claisen process,2 furnishing
significant amounts of the corresponding 3-methyl benzofuran
(16) and 3-methyl indole (17) products (Figure 3). Apparently,
when silver(I) and platinum(II) salts are combined, the catalytic
process promoted by platinum overrides that offered by silver,
and regiospecific insertion of the aryl-hydrogen bond across
the triple bond of the propargyl group is observed.
while depropargylation was limited to less than 3% and the
levels of ketones 15a and 15b (attributed to hydration of the
activated alkyne) did not exceed 2%.11 The content of residual
gold in the isolated chromenes was approximately 200 ppm.
This was not regarded as an issue since heavy metal levels are
managed effectively over the remaining four steps of the
synthesis.2 The low catalyst loading and the ease of synthesis
of substrates 13a and 13b offset the cost of triphenylphosphine-
gold(I) bis(trifluoromethanesulfonyl)imidate and render this
route superior to the Claisen approach on the grounds of both
operability and cost.
2.3.2. Intramolecular Friedel-Crafts Approaches. Follow-
ing our initial brainstorm, one of the most attractive approaches
involved the intramolecular Friedel-Crafts approach via activa-
tion of appropriately substituted aryl alkyl ether substrates. In
order to explore this approach we prepared benzyl 3-aryloxy-
acrylate 18 in 78% yield (cis/trans ∼1:6) via the DABCO-
catalysed addition of 12b to benzyl propiolate (Scheme 5).12
The subsequent one-pot hydrogenolysis/hydrogenation to give
the desired substrate 19 was more problematic than originally
anticipated. Most of the commercial Pd/C catalysts were
successful in the hydrogenolysis of the benzyl ester, but very
few were effective in the hydrogenation of the 3-aryloxyacrylic
acid intermediate. At best, hydrogenation of the double bond
took no less than one week to reach completion (THF, 50 °C,
50 psi). In addition, benzyl propiolate itself is not commercially
available and had to be prepared by treating propiolic acid with
benzyl trichloroacetimidate. Interestingly, only 11% of the
combined mass of these reagents ends up as part of the structure
in 19, rendering the atom economy of the process extremely
poor. In contrast, we succeeded in preparing 19 in one step
and 88% yield, by reacting the phenolate anion of 12b with
ꢀ-propiolactone (Scheme 5).13
Figure 3. Impurities from Ag(I) catalysis are also associated
with the Claisen route.
The combination of chloro(triphenylphospine)gold(I) and
silver(I) triflate generates the triphenylphospine-gold(I) cation,
presumably the active species, but triphenylphospine-gold(I)
triflate is not an isolable precursor of the active form. Triph-
enylphosphine gold(I) bis(trifluoromethanesulfonyl)imidate
[(Ph3P)AuNTf2] (Gagozs’ catalyst) is commercially available
and does not require additives to generate the active species.8b
The work using this catalyst was performed on small scale (up
to 5 g) due to its limited commercial availability. Under our
optimised conditions (0.1 mol % of catalyst, toluene, 85 °C),
the desired chromenes 14a and 14b were isolated in 80% yield,
With a good supply route to 19 in hand, we explored
activation of the carboxylic acid terminus and promotion of the
intramolecular Friedel-Crafts reaction. Treatment of a toluene
(10) Shen, R.; Zhu, S.; Huang, X. J. Org. Chem. 2009, 74, 4118.
(11) The levels of the ketone impurity may be connected to the difficulty
in rigorously excluding water on the small scale at which these
experiments were conducted.
(12) Fan, M.-J.; Li, G.-Q.; Li, L.-H.; Yang, S.-D.; Liang, Y.-M. Synthesis
2006, 14, 2286.
(13) Buckle, D. R.; Eggleston, D. S.; Houge-Frydrych, C. S. V.; Pinto,
I. L.; Readshaw, S. A.; Smith, D. G.; Webster, R. A. B. J. Chem.
Soc., Perkin Trans. 1 1991, 2763.
94
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Scheme 5. Syntheses of the Friedel-Crafts substrate 19a
Scheme 6. Benzopyranone synthesis via intramolecular
Friedel-Crafts cyclisationa
a Reagents and conditions: a) Toluene, (Cl3CCO)2O or Me3CCOCl/ Et3N, rt;
b) BF3 · Et2O, toluene 99 °C.
Scheme 7. Synthesis of the bromobenzopyran intermediate
25a
a Reagents and conditions: a) Benzyl trichloroacetimidate, DCM, rt; b) 12b,
DABCO (5 mol %), DCM, rt; c) Pd/C, THF, 50 °C, 50 psi; d) tBuOK, 2-MeTHF,
rt.
solution of 19 at 60 °C with trichloroacetic anhydride generated
the corresponding mixed anhydride 20a as a crystalline solid
suspended in the solution (Scheme 6). Addition of a catalytic
amount of BF3 ·Et2O to this suspension at 100 °C gave the crude
benzopyranone 21 in 52% isolated yield but in moderate
purity.14 By studying the activation and cyclisation events in
isolation we established that the performance of the Friedel-
Crafts reaction reflected the quality of 20a. When the mixed
anhydride was preisolated by filtration of the original slurry and
subjected to the cyclisation conditions, the desired benzofura-
none was obtained in excellent quality at 60% yield. Isolation
of the mixed anhydride 20a was problematic due to its
hygroscopicity, and this was compounded by contamination
with the symmetrical anhydride 20b, even when large amounts
of trichloroacetic anhydride were used. Depending on the
solvent and the mode of addition of trichloroacetic acid
anhydride to 19, the ratio between 20a and 20b ranged from
a Reagents and conditions: a) NaBH4, MeOH, rt; b) H2SO4 (conc.), MeOH,
Pd/C, H2, 50 °C, 50 psi; c) NBS, MeOH, rt.
intermediate 21 from phenol 12b in 67% yield through a
simplified, telescoped process.
1
The synthesis of the desired benzopyran skeleton could now
be completed by reduction of the benzopyranone carbonyl to
the methylene group. Unfortunately, ketone 21 was inert to a
wide range of hydrogenation conditions that encompassed
solvents, pressures, temperatures, additives, and no less than
20 catalysts. Our substrate was also inert to trifluoroacetic acid/
triethyl silane, while significant decomposition occurred using
a modification of the Clemmensen reduction.15 Exploring a
stepwise approach, addition of half an equivalent of sodium
borohydride to a methanolic suspension of 21 at ambient
temperature resulted in the rapid and quantitative reduction of
the benzopyranone to the corresponding benzopyranol 22 which
was isolated in greater than 99% yield and excellent quality
(Scheme 7). We subsequently established that reduction of the
secondary alcohol 22 to benzopyran 23 by hydrogenation is
facilitated by the presence of a strong acid. Upon addition of
sulphuric acid, 22 converts into two intermediates (initially ∼1:1
by HPLC area) both of which convert to 23 when the catalyst
is added and a hydrogen atmosphere is established. These
intermediates were identified as chromene 14b and enol ether
5:1 to 1:10 as determined by H NMR spectroscopy. Higher
levels of symmetrical anhydride 20b (which did not cyclise
under the reaction conditions) impacted adversely on the
intramolecular Friedel-Crafts reaction, at least when BF3 ·Et2O
was used as the Lewis acid. We reasoned that the high reactivity
of 20a was responsible for both moisture sensitivity and
formation of 20b. In order to address these issues we decided
to explore pivaloyl chloride as the activator of 19 as it would
lead to a more stable activated species. Rewardingly, treatment
of 19 with pivaloyl chloride afforded 20c quantitatively and
without contamination from 20b. The pivaloyl-mixed anhydride
is a crystalline solid which can be stored for several days under
ambient conditions. Exposure of this material to a catalytic
amount of BF3 ·Et2O in toluene for 7 h at 100 °C provided 21
as a yellow crystalline solid in excellent quality and in 83%
yield from 19 (Scheme 6).
Subsequently, we demonstrated that isolation of 19 and 20c
was not necessary as we prepared the key benzopyranone
(14) (a) Wilkinson, M. C.; Saez, F.; Hon, W. L. Synlett 2006, 7, 1063. (b)
Bianco, G. G.; Ferraz, H. M. C.; Costa, A. M.; Costa-Lotufo, L. V.;
Pessoa, C.; de Moraes, M. O.; Schrems, M. G.; Pfaltz, A.; Silva, L. F.,
Jr. J. Org. Chem. 2009, 74, 2561.
(15) Prugh, J. D.; Alberts, A. W.; Deana, A. A.; Gilfillian, J. L.; Huff,
J. W.; Smith, R. L.; Wiggins, J. M. J. Med. Chem. 1990, 33, 758.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
95

24 (major product under nonreductive conditions), presumably
arising from dehydration of 22 and isomerisation of 14b,
respectively. Enol ether 24 is hydrogenated at a much slower
rate to give 23 although conversion to the desired benzopyran
via equilibration to 14b cannot be dismissed. The rate of their
hydrogenation was greatly dependent on the type of catalyst
used and best results were obtained with unreduced palladium
on wet charcoal. When other types of catalysts were used, 24
persisted for days. Under our optimised conditions 23 was
isolated in approximately 90% yield and excellent quality. When
ethanol was used as the solvent in the reductive steps, significant
trans-esterification of the methyl ester group in 22 and 23 was
observed. For this reason, methanol was chosen as the reaction
solvent, which also allowed the borohydride reduction to be
telescoped into the hydrogenation process.
Methanol was also found to be the optimum solvent in the
bromination of 23 to 25 with NBS (equally clean although much
slower in EtOAc, THF, or TBME at ambient temperature)
(Scheme 7). This aspect rendered the work-up of the combined
reduction steps and isolation of 23 unnecessary, as we dem-
onstrated the synthesis of 25 from 21 in a one-pot, telescoped
process in 67% yield on a 20 g scale (Scheme 8).
sation. The two routes are similar in length, cost, and efficiency
and constitute attractive alternatives to previously established
routes.16
3. Experimental Section
Phenol 12b. Methyl 4-[(2,2-Dimethylpropanoyl)amino]-2-
hydroxybenzoate. 4-Amino salicylic acid (700 g, 4.57 mol, 1
equiv) was dissolved in methanol (5.25 L, 7.5 vol), and the
clear brown solution was cooled to 0-5 °C. Sulphuric acid (1.4
L, 2 vol, 2.57 kg, 26.2 mol, 5.73 equiv) was added slowly,
maintaining the temperature below 20 °C. The reaction was
then heated to 80 °C and kept at that temperature for 3 h before
cooling to 20-30 °C. Water (10.5 L, 15 vol) was added, and
the pH was adjusted to 7 using 10 M NaOH (3.0 L, 4.29 vol,
9 °C exotherm). The white precipitate was filtered under
vacuum, washed with water (1.4 L, 2 vol), and dried for 2 days
under vacuum at 80 °C to give dry 11b (727 g, 1.04 wt, 1.04
mol, 95.1% yield). In another experiment, the wet cake obtained
after the water wash was dissolved directly in ethyl acetate (10.5
L, 15 vol) and added to water (3.5 L, 5 vol). Sodium hydrogen
carbonate (525 g, 0.75 wt, 6.25 mol, 1.37 equiv) was added
followed by pivaloyl chloride (700 g, 1 wt, 5.8 mol, 1.27 equiv),
and the reaction mixture was stirred at 25-30 °C for approxo-
mately 2 h. After settling of the biphase, the aqueous phase
was removed, and the organic phase was washed with water
(2 × 3.5 L, 2 × 5 vol). The organic phase was then concentrated
to approximately 2.8 L (4 vol) and cooled to 20 °C; heptane (7
L, 10 vol) was then added. The resulting slurry was filtered,
and the cake was washed with heptane (2.1 L, 3 vol) and dried
under vacuum at 50 °C to give the titled compound (954 g,
1.31 wt, 3.8 mol, 83% yield). 1H NMR (400 MHz, CDCl3) δ
10.83 (s, 3H), 7.77 (d, J ) 8.80 Hz, 1H), 7.37 (br s, 1H), 7.18
(d, J ) 2.20 Hz, 1H), 7.13 (dd, J ) 8.68, 2.08 Hz, 1H), 3.93
(s, 3H), 1.33 (s, 9H); 13C NMR (100 MHz, CDCl3) δ 177.4,
170.2, 163.0, 144.6, 131.1, 110.5, 107.8, 106.45, 52.1, 40.2,
27.5; HRMS calculated for the protonated molecular ion
(MH+): C13H18NO4 252.1236, found 252.1226.
Scheme 8. Summary of the successful syntheses of the
benzopyran intermediatea
Propargyl Ether 13b. Methyl 4-[(2,2-Dimethylpropano-
yl)amino]-2-(2-propyn-1-yloxy)benzoate. Potassium carbonate
(2.00 g, 14.5 mmol, 1.22 equiv) was added to a solution of
methyl 4-[(2,2-dimethylpropanoyl)amino]-2-hydroxybenzoate
(3.00 g, 11.9 mmol, 1 equiv) in acetone (15 mL, 5 vol) and the
resulting slurry heated at reflux for 20 h. The mixture was
cooled, diluted with water (75 mL, 25 vol), and extracted with
ethyl acetate (30 mL, 10 vol). The organic extract was washed
with water (30 mL, 10 vol) and then 10% w/w sodium chloride
in water, was dried over sodium carbonate, and concentrated
in Vacuo. The residue was suspended in toluene (9 mL, 3vol)
and the mixture heated until the solid had dissolved. The
solution was cooled to ambient temperature. The crystalline
solid was collected under suction, washed with toluene (3 mL,
1 vol) followed by heptane (6 mL, 2 vol), and dried in Vacuo
at 45 °C to afford 13b (3.10 g, 1.03 wt, 10.7 mmol, 90%); 1H
NMR (400 MHz, CDCl3) δ 7.84 (d, J ) 8.31 Hz, 1H), 7.76
a Reagents and conditions: a) propargyl bromide, K2CO3, acetone, reflux; b)
0.1 mol % (Ph3P)AuNTf2 ·0.5C7H8, toluene, 85 °C; c) tBuOK, ꢀ-propiolactone,
2-MeTHF, rt; d) Me3CCOCl, Et3N, BF3 · Et2O, toluene 100 °C; e) (i) NaBH4,
MeOH, rt, 99%. (ii) H2SO4 (conc.), MeOH, Pd/C, H2, 50 °C, 50 psi, 90%. (iii)
NBS, MeOH, rt, 75% unoptimised.
By-passing the isolation of 25 and further telescoping the
protected bromobenzopyran intermediate into the hydrolysis of
the ester and pivaloyl groups is also being considered.
To conclude, three pivotal disconnections to a key benzopy-
ran intermediate were considered, and a number of synthetic
routes were evaluated experimentally. The disconnection across
the bond joining the aromatic ring to the benzylic benzopyran
carbon atom proved the best approach as it produced two
successful syntheses (Scheme 8). The first route involves the
use of a transition-metal catalysed cycloisomerisation, whereas
the other is based on an intramolecular Friedel-Crafts cycli-
(16) (a) Ahmed, M.; Miller, N. D.; Moss, S. F.; Sanger, G. J. WO
2007096352, 2007. (b) Ishii, H.; Ishikawa, T.; Takeda, S.; Ueki, S.;
Suzuki, M.; Harayama, T. Chem. Pharm. Bull. 1990, 38, 1775. (c)
Kakigami, T.; Baba, K.; Usui, T. Heterocycles 1998, 48, 2611–2619.
(d) Baba, Y.; Usui, T.; Kakiue, T.; Baba, K. JP 08157466, 1996.
96
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

(d, J ) 1.96 Hz, 1H), 7.43 (br s, 1H), 6.95 (dd, J ) 8.56, 1.96
Hz, 1H), 4.82 (d, J ) 2.45 Hz, 2H), 3.87 (s, 3H), 2.55 (t, 2.45
Hz, 1H), 1.33 (s, 9H); 13C NMR (100 MHz, CDCl3) δ 177.17,
165.72, 158.68, 143.13, 132.92, 115.91, 111.57, 105.08, 78.15,
75.94, 56.88, 51.84, 39.81, 27.29; HRMS calculated for the
protonated molecular ion (MH+): C16H20NO4 290.1392, found
290.1395 (APCI).
HRMS calculated for the protonated molecular ion (MH+):
C16H21NO5 calculated 324.1447, found 324.1448.
Benzopyranone 21. Methyl 5-[(2,2-Dimethylpropanoyl)ami-
no]-4-oxo-3,4-dihydro-2H-chromene-8-carboxylate. To a solu-
tion of 19 (11 g, 0.034 mol, 1 equiv) in 2-MeTHF (80 mL,
7.27 vol)spreisolated or telescoped from the previous steps
was added triethylamine (3.77 g, 0.037 mol, 1.09 equiv). The
new solution was added dropwise to a solution of pivaloyl
chloride (12.5 g, 0.104 mol, 2.94 equiv) in toluene (80 mL,
7.27 vol) at 20 °C. The precipitated triethylammonium chloride
was removed by filtration, and the solution was solvent-swapped
to pure toluene. Further put and take distillation with toluene
was performed to remove the excess of pivaloyl chloride. The
toluene suspension (solution at >60 °C) of the mixed anhydride
was diluted with toluene (90 mL, 8.18 vol) and heated to 100
°C. A catalytic amount of BF3 ·Et2O (0.38 g, 8 mol %) was
added, and the reaction mixture was kept at the same temper-
ature for at least 7 h before it was cooled to 60 °C. The reaction
mixture was quenched with saturated sodium bicarbonate
solution (45 mL, 4 vol). The aqueous phase was discarded, and
the organic phase was washed with 2 M HCl (45 mL, 4 vol)
and finally with water (45 mL, 4 vol) (all washes conducted at
>60 °C). The toluene solution was concentrated to approxi-
mately 35 mL and cooled to ambient temperature, and heptane
(60 mL, 5.45 vol) was added. The resulting slurry was filtered,
washed with heptane (30 mL, 2.72 vol, and dried to afford 21
Chromene 14b. Methyl 5-[(2,2-Dimethylpropanoyl)amino]-
2H-chromene-8-carboxylate. A mixture of aryl propargyl ether
13b (731 mg, 2.53 mmol) and (Ph3P)AuNTf2 ·0.5C7H8 (2.0 mg,
2.5 µmol, 0.1 mol %) in toluene (14.6 mL, 20 vol) was stirred
at 85 °C under nitrogen for 2 h. The reaction mixture was
allowed to cool to 55 °C and was filtered and washed with
toluene (3.6 mL, 5 vol). The filtrate was concentrated in Vacuo
and the residual solid dissolved in hot toluene (2.9 mL, 4 vol).
The solution was allowed to cool and stirred at room temper-
ature overnight. The crystalline solid was collected under
suction, washed with toluene (1.5 mL, 2 vol), and dried in Vacuo
at 45 °C to afford 14b (584 mg, 0.8 wt, 2.02 mmol, 80% yield)
1
as an off-white solid; mp 153-155 °C (dec); H NMR (400
MHz, CDCl3) δ 7.69 (d, 1H, J ) 8.8 Hz); 7.42 (d, J ) 8.8 Hz,
1H), 7.37 (br s, 1H), 6.38 (m, 1H), 5.96 (dt, J ) 9.8, 3.9 Hz,
1H), 4.83 (m, 2H), 3.87 (s, 3H), 1.35 (s, 9H); 13C NMR (100
MHz, CDCl3) δ 176.81, 165.83, 155.07, 136.72, 131.49, 122.71,
118.97, 115.96, 115.65, 115.36, 64.97, 51.93, 39.82, 27.68;
HRMS calculated for the protonated molecular ion (MH+):
C16H20NO4 290.1392, found 290.1387 (APCI). The product
contained ∼220 ppm Au and 6 ppm P by ICP. HPLC analysis
of the combined filtrate and wash indicated a ∼10% yield of
14b was present by comparison to a standard solution.
1
(8.6 g, 0.78 wt, 0.28 mol, 83% yield). H NMR (400 MHz,
DMSO-d6) δ 12.36 (s, 1H), 8.29 (d, J ) 9.05 Hz, 1H), 7.96 (d,
J ) 9.05 Hz, 1H), 4.61 (t, J ) 6.60 Hz, 2H), 3.80 (s, 3H), 2.94
(t, J ) 6.60 Hz, 2H), 1.29 (s, 9H). 13C NMR (100 MHz,
DMSO-d6) δ 196.90, 177.68, 164.53, 161.88, 145.09, 138.79,
113.52, 110.25, 108.58, 66.41, 51.87, 39.97, 37.44, 27.04.
HRMS calculated for the protonated molecular ion (MH+):
C16H19NO4 calculated 306.1341, found 324.1334.
3-Aryloxy Propionic Acid 19. 3-({5-[(2,2-Dimethylpro-
panoyl)amino]-2-[(methyloxy)carbonyl]phenyl}oxy) Propanoic
Acid. To a solution of 12b (20 g, 0.08 mol) and ꢀ-propi-
olactone (5 mL, d ) 1.15, 0.08 mol) in 2-MeTHF (250
mL, 12.5 vol) at 20 °C, was added a 20% w/w solution
of KOtBu in THF (45 mL, 0.08 mol) over 5 min, and the
reaction mixture was stirred at the same temperature for
at least 4 h (marginal improvement in conversion by HPLC
after reacting overnight). The reaction mixture was
quenched by the addition of water (120 mL, 6 vol), and
the pH of the aqueous phase was adjusted from 14 to 8-9
with 2 M HCl. The aqueous phase was separated, and its
pH was adjusted to 1 with 2 M HCl before it was extracted
with 2-MeTHF (160 mL, 8 vol). The product in the
organic phase is of sufficient purity to be telescoped into
the next stage directly (after concentration to 80 mL).
Alternatively, the 2-MeTHF solution may be dried over
sodium sulphate and evaporated to dryness to isolate 19
directly or solvent-swapped to toluene to give a slurry of
19 which is then filtered to afford the titled compound
Benzopyran 23. Methyl 5-[(2,2-Dimethylpropanoyl)amino]-
3,4-dihydro-2H-chromene-8-carboxylate. Sodium borohydride
(0.3 g, 8 mmol, 0.51 equiv) was added to a suspension of
benzopyranone 21 (5.0 g, 0.015 mol, 1 equiv) in methanol (50
mL, 10 vol) at 20 °C, and the resulting solution was stirred at
the same temperature for at least 15 min (reduction to the
alcohol is complete) before it was quenched with glacial acetic
acid (5 mL, 1 vol). Concentrated sulphuric acid (2.5 mL, 0.05
mol, 0.5 vol) was added followed by 5% Pd/C (5.0 g, 2.4 mmol,
0.29 equiv unreduced catalyst E196 R/W by Degussa). A
hydrogen atmosphere (50 psi) was established, and the solution
was heated to 50 °C and stirred for at least 24 h. The reaction
mixture was filtered over Celit, and the Celite cake was washed
with methanol (10 mL, 2 vol) which was combined with the
filtrate (this solution may be processed further to give 25 by
treatment with NBS). The methanolic solution was concentrated
and solvent-swapped to toluene by distillation (∼100 mbar).
The resulting toluene solution (∼50 mL) was washed succes-
sively at 60 °C with water (20 mL) and aqueous 20% KHCO3
solution (2 × 50 mL). The toluene solution was then concen-
trated to approximately 10 mL, resulting in a slurry of the
product which was filtered, washed with toluene (10 mL, 2 vol),
and dried overnight under vacuum at 60 °C, to afford the titled
compound (4.25 g, 0.85 wt, 14.5 mmol, 89% yield). 1H NMR
1
(22 g, 1.1 wt, 0.68 mol, 88% yield). H NMR (400 MHz,
DMSO-d6) δ 9.42 (s, 1H), 7.67 (d, J ) 8.56 Hz, 1H),
7.60 (d, J ) 1.71 Hz, 1H), 7.42 (dd, J ) 8.56, 1.71 Hz,
1H), 4.19 (t, J ) 6.24 Hz, 2H), 3.74 (s, 3H), 2.74 (t, J )
6.24 Hz, 2H), 1.25 (s, 9H). 13C NMR (100 MHz, DMSO-
d6) δ 177.04, 172.01, 165.44, 158.47, 144.54, 131.83,
113.97, 111.29, 104.60, 64.55, 51.37, 39.14, 34.00, 27.02.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
97

(400 MHz, DMSO-d6) δ 8.87 (s, 1H), 7.48 (d, J ) 8.31 Hz,
1H), 6.92 (d, J ) 8.31 Hz, 1H), 4.17 (m, 2H), 3.78 (s, 3H),
2.63 (t, J ) 6.48 Hz, 2H), 1.92 (m, 2H), 1.26 (s, 9H). 13C NMR
(100 MHz, DMSO-d6) δ 176.30, 165.95, 154.54, 140.75,
127.98, 119.29, 117.14, 116.23, 65.89, 51.64, 39.82, 27.26,
20.72, 20.58. HRMS calculated for the protonated molecular
ion (MH+): C16H22NO3 292.1549, found 292.1537.
Bromobenzopyran 25. Methyl 6-Bromo-5-[(2,2-dimethyl-
propanoyl)amino]-3,4-dihydro-2H-chromene-8-carboxylate. To
a solution of benzopyran 23 (2.12 g, 7.26 mmol, 1 equiv) in
methanol (15 mL, 6.91 vol) (preisolated or telescoped from the
previous step after removal of the catalyst) was added N-
bromosuccinimide (1.3 g, 7.26 mmol, 1 equiv) in a single charge
at 20 °C. The reaction mixture was stirred for at least 15 min
before it was concentrated and solvent-swapped to toluene by
distillation (∼100 mbar). The resulting toluene solution (∼20
mL) was washed at 60 °C with aqueous 20% KHCO3 solution
(2 × 10 mL). The toluene solution was then concentrated to
dryness to afford 1.8 g, 59% yield of sufficiently pure (>98%
by 1H NMR) bromobenzopyran 25 to be processed further. 1H
NMR (400 MHz, DMSO-d6) δ 9.18 (s, 1H), 7.72 (s, 1H), 4.17
(t, J ) 4.89 Hz, 2H), 3.80 (s, 3H), 2.58 (m, 2H), 1.92 (m, 2H),
1.26 (s, 9H). 13C NMR (100 MHz, DMSO-d6) δ 175.92, 164.77,
153.62, 139.47, 130.68, 124.89, 119.49, 112.63, 66.06, 52.09,
38.75, 27.25, 21.56, 20.49. HRMS calculated for the protonated
molecular ion (MH+): C16H21BrNO3 370.0654, found 370.0644.
Received for review July 23, 2009.
OP900188V
98
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development