Scalable synthesis of 8-amino-3-hydroxy-6 H -benzo[ c ]chromen-6-one: Key intermediate for SEGRA via the Hurtley reaction

Article abstract of DOI:10.1021/op500334b

A practical and scalable process for the preparation of 8-amino-3-hydroxy-6H-benzo[c]chromen-6-one in multihundred kilogram amounts has been developed. The key features of this synthesis are the application of the Hurtley reaction with a copper and base combination and the development of a purification process. The new synthesis improved the total yield from 49.0% to 59.5% and reduced the number of steps from three to two. Compared with the conventional medicinal route, manufacturing costs were reduced significantly by the use of inexpensive, easy to procure materials.

Full text of DOI:10.1021/op500334b

Communication
pubs.acs.org/OPRD
Scalable Synthesis of 8‑Amino-3-hydroxy‑6H‑benzo[c]chromen-6-
one: Key Intermediate for SEGRA via the Hurtley Reaction
Kazuhiro Kudo*,†,‡ and Noriyoshi Yamamoto†
†Nara Research & Development Center, Santen Pharmaceutical Co., Ltd, 8916-16, Takayama-cho, Ikoma-shi, Nara 630-0101, Japan
‡Product Supply Division, Santen Pharmaceutical Co., Ltd, 348-3, Aza Suwa Oaza Shide Taga-cho Inukami-gun, Shiga 522-0314,
Japan
In the conventional medicinal route (Scheme 1),3c a number
ABSTRACT: A practical and scalable process for the
preparation of 8-amino-3-hydroxy-6H-benzo[c]chromen-
6-one in multihundred kilogram amounts has been
developed. The key features of this synthesis are the
application of the Hurtley reaction with a copper and base
combination and the development of a purification
process. The new synthesis improved the total yield
from 49.0% to 59.5% and reduced the number of steps
from three to two. Compared with the conventional
medicinal route, manufacturing costs were reduced
significantly by the use of inexpensive, easy to procure
materials.
of SEGRA compounds were synthesized from key intermediate
1a. Synthesis of 1a was conducted in a three-step procedure
that involved the Suzuki−Miyaura coupling of commercially
available 2 and 3 in quantitative yield. Cleavage of the methyl
ether of 4 with BBr3 and successive intramolecular cyclization
provided 5, which has 3-hydroxy-6H-benzo[c]chromen-6-one
framework, in 51% yield. The nitro moiety of 5 was reduced to
give 1a in 96% yield. In this synthesis, 1a was obtained in three
steps with a total yield of 49%. However, manufacturing costs
in this process are extremely high, and there are serious
environmental concerns due to the use of 5 mol % of palladium
catalyst and the extensive use of boronic acid 2 and BBr3/
CH2Cl2.4a,b
Therefore, a more economical synthetic route for 1a is
desirable.
INTRODUCTION
■
In synthetic approaches for the key intermediate 1a, the
Hurtley reaction, which is a one-pot condensation of phenols
and 2-halobenzoic acid in the presence of a copper catalyst, is
considered to be appropriate.5 Although resorcinol 7 and
copper catalyst would be cost-effective and the number of steps
would be reduced, few examples of the reaction of nitro or
amino analogues on the 5-position of bromobenzoic acid 6
have been reported, and there have been no reports of such a
reaction with high yield (Scheme 2).6 In this study, we
investigated a practical and scalable synthesis of 1a via the
Hurtley reaction.
Glucocorticoids (GC) play a dominant role in the treatment of
inflammatory conditions. Apart from their potent anti-
inflammatory activity, long-term, and high dose treatment
may lead to serious adverse effects. It has been hypothesized
that transrepression (TR) is the basis of their anti-inflammatory
effects, whereas transactivation (TA) has been assumed to
cause their side effects.1
In recent years, a selective glucocorticoid receptor agonist
(SEGRA) that can separate TR and TA activities without a
steroid framework was reported.2 SEGRA compounds (Figure
1) exert selective TR and exhibit a similar GR binding potency
RESULTS AND DISCUSSION
■
To investigate feasibility, the Hurtley reaction was assessed in
the reaction of 5-amino-2-bromobenzoic acid 6a (Table 1). A
reaction of 6a was found to give trace amounts of 1a in 30 min
(entry 1). A significant number of byproducts occurred with the
use of a free amino substituent; consequently, it was considered
that improving the yield would be difficult. It was found that
acetyl protection of the amino substituent to give 6b was
affordable and that this compound was readily available in large
quantities. When the reaction with 6b was conducted, 1b was
isolated by filtration (39% yield) with fewer byproducts than in
entry 1 (entry 2). When the amount of reagents was increased
(3.0 equiv of 7, 3.3 equiv of NaOH, and 0.3 equiv of CuSO4)
the yield improved to 50% (entry 3). For temperature
screening at 40 and 60 °C, a superior yield was obtained at
60 °C (entry 4, 5). When the reaction was performed with a
Figure 1. General structure of SEGRA compounds.
to that of dexamethasone in vitro. In an in vivo assay, SEGRA
compounds showed a remarkable anti-inflammatory effect after
topical dosing to the eye and side effect dissociation.3 To
evaluate efficacy in preclinical studies and clinical trials,
multikilogram quantities of SEGRA compounds were required.
Consequently, the development of a reliable and cost-effective
synthesis process with high yield was important.
Received: October 25, 2014
© XXXX American Chemical Society
A
DOI: 10.1021/op500334b
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
Scheme 1. Initial medicinal route for key intermediate 1a
Scheme 2. New synthetic route of 1 by the Hurtley reaction
Table 1. Optimization of the Hurtley reaction of 6a, 6b
Table 2. Optimization of the copper catalyst and main
impurities
yield of
yield of 1b
ab
yield of
c
yield of
B (%)
yield of
C (%)
temp time resorcinol NaOH CuSO4
1a, 1b
,
ab
,
entry
catalyst
(%)
65
A (%)
entry
R
(°C)
(h)
7 (equiv) (equiv) (equiv)
(%)
d
c
1
2
3
4
5
CuSO4
13
13
15
16
10
16
16
17
16
12
3
6
5
5
2
1
2
3
4
5
6
7
8
9
10
H
100
100
100
60
0.5
0.5
1
2.0
2.0
3.0
3.0
3.0
2.0
3.0
3.0
3.0
3.0
2.2
2.2
3.3
3.3
3.3
3.3
2.0
3.3
3.3
3.3
0.10
0.10
0.3
trace
CuBr
58
57
60
67
Ac
39
50
65
44
54
32
41
44
64
Cu(OAc)2
CuCl2
CuI
1
0.3
40
3
3.0
60
2
0.3
60
5
0.3
60
3
0.05
0.1
60
3
60
1.5
0.5
a
Isolated yield after filtration and washing with 6 M HCl aq, H2O, and
a
Isolated yield after filtration and washing with 6 M HCl aq and H2O.
b
EtOH. All yields excluded resorcinol 7 that was not purified; no other
b
All yields excluded resorcinol 7 that was not purified; no other
c
impurities were confirmed by 1H NMR. After reaction, a trace
c
impurities were confirmed by 1H NMR. The yields of impurities were
confirmed by HPLC assay in the filtrate after filtering off precipitated
amount of target materials was observed in crude material by 1H
NMR.
d
target material 1b. This entry was referred from entry 3 in Table 1.
decreased amount of resorcinol 7 and NaOH at 60 °C, the
yield decreased (entry 6, 7). When screening for the amount of
CuSO4, a lower equivalent resulted in low yield (entry 8, 9),
and increasing the amount did not improve the yield (entry
10). The deactivation of the copper catalyst in lower
equivalents was considered. It was thought that activity may
have been suppressed by the generated solid, and consequently,
the activity was reduced as the reaction progressed.
Further investigation was undertaken to determine how the
copper counterion effect of the copper catalysts affected yield
and byproducts (A, B, and C),7 as shown in Table 2. It was
found that the instability and low yield of the Hurtley reaction
primarily arose from the production of the byproducts. When
the reaction was performed using CuBr, Cu(OAc)2, and CuCl2
(entry 2, 3, 4), the yields were slightly lower than when CuSO4
(entry 1) was used with similar byproduct profiles. CuI resulted
in a slightly higher yield as a consequence of the reduced
generation of all byproducts (entry 5). In this screening, the
copper counterions did not affect the yield significantly;
however, the copper counterion may have influenced the
ratio of byproducts.
The results of base screening are presented in Table 3.
Compared to NaOH, KOH and LiOH (entry 1, 2) resulted in
almost equal yields with similar byproduct profiles. Despite
being weak bases, K2CO3 and Na2CO3 improved the yield
slightly (entry 3, 4). Hydroxide bases inhibited the production
B
DOI: 10.1021/op500334b
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
Table 3. Optimization of the base
Table 4. Optimization of the base and copper catalyst of the
Hurtley reaction
yield of
yield of yield of yield of remaining
ab
,
entry
base
KOH
1b (%)
A (%)
B (%)
C (%)
6b (%)
yield
yield
of
yield
of
yield
of
temp
entry (°C)
of 1b
remaining
6b (%)
1
2
3
4
5
6
7
61
59
65
67
48
59
43
14
13
3
17
19
7
5
ab
,
base
(%)
A (%) B (%) C (%)
LiOH
5
1
2
3
4
60
50
40
50
Na2CO3
Na2CO3
Na2CO3
73
82
79
82
6
5
4
7
7
6
6
5
4
K2CO3
Na2CO3
Li2CO3
NaHCO3
Et3N
11
13
15
10
11
c
5
6
2
11
9
13
11
25
2
6
7
c
Na2CO3/
NaOH
14
c
5
6
7
8
9
50
50
50
50
50
K2CO3
78
47
50
73
55
4
6
6
a
Isolated yield after filtration and washing with 6 M HCl aq and H2O.
Li2CO3
Cs2CO3
NaHCO3
KHCO3
10
4
15
9
7
b
All yields excluded resorcinol 7 that was not purified; no other
13
7
15
4
1
impurities were confirmed by H NMR except for 6b.
5
7
6
14
17
a
Isolated yield after filtration and washing with 6 M HCl aq and H2O
of byproduct C, and carbonate bases inhibited the production
of A and B but did not inhibit the production of C. Li2CO3
(entry 5) afforded lower yields but resulted in the production of
greater amounts of byproducts A and B than those observed for
K2CO3 and Na2CO3. NaHCO3 showed an approximately
equivalent yield with comparable production of byproducts A,
B, and C and the remaining starting material (entry 6). Et3N
showed relatively low yields with relatively large amounts of
byproducts A and B and the remaining starting material (entry
7).
From the results of the investigation to this point, it was
found that the Hurtley reaction was able to afford the target
material in approximately 65% yield by protecting the amino
function with an acetyl group and by increasing the amount of
reagents. To achieve further improvement, we focused on
profiling the byproducts using different copper and base
combinations. Yield was not improved when carbonate bases
were used; however, the ratios of byproducts A and B were
reduced. With CuI, a drastic reduction of byproduct C was
observed. Given these results, combined use of a carbonate base
and CuI was explored.
Table 4 shows the results of the Hurtley reaction with CuI
and a carbonate base. At 60 °C, 73% yield was obtained with
CuI and Na2CO3 (entry 1), and all byproducts were reduced.
At 50 °C, the yield was greater than 80% (entry 2), and at 40
°C the yield slightly decreased, and a small amount of starting
material remained (entry 3). It was found that the initial
addition of 1 equiv of NaOH aq to bromobenzoic acid to
obtain the corresponding sodium carboxylate and the addition
of 2.3 equiv of Na2CO3 later in the reaction resulted in an assay
yield of 82% (entry 4). These separate additions avoided the
foaming that is generated by mixing benzoic acid and a
carbonate base, which is a consideration for a scaled-up
synthesis. K2CO3 resulted in a slightly lower assay yield as
compared with case when Na2CO3 (entry 5) is used. There is
no significant difference between K2CO3 and Na2CO3;
however, Na2CO3 is more favorable as it has a lower molecular
weight. Li2CO3 resulted in low yield and increased byproducts
A and B (entry 6). Cs2CO3 also resulted in low yield, increased
C, and 15% remaining starting material (entry 7). NaHCO3
resulted in relatively low yield (entry 8) with relatively low
ratios of byproducts. Use of KHCO3 decreased the yield and
showed increased amounts of B and C (entry 9).
b
and no further purification to remove CuI. All yields excluded
resorcinol 7 that was not purified; no other impurities were confirmed
by H NMR except for 6b. HPLC assay yield.
c
1
Possible reasons for the high yield resulting from the
combination of CuI and a carbonate base were considered.
Here it was thought that a weak base was not sufficient to
achieve carbon−carbon and carbon−oxygen bond formation in
this reaction. Consequently, the addition of the phenolic
hydroxyl group on resorcinol 7 and hydroxide to the substrate
was thought to result in reduced amounts of A and B and lead
to increased decomposition of the bromide, which resulted in
C. In contrast, monovalent CuI had a strong potential to form
carbon−carbon and carbon−oxygen bonds,8 and consequently,
the ratio of decomposition was decreased. It was assumed that,
compared with conventional the Hurtley reactions, the
combination of CuI and a carbonate base improved catalytic
activity and reduced byproducts.
To achieve effective synthesis of SEGRA compounds, the
removal of copper was important because it was found that
residual copper (>900 ppm) showed a critical influence in the
subsequent step.9 However, removing copper was found to be
difficult in a large-scale process. At a small scale (100−200 g),
washing the target material, which was filtered after the
reaction, in an aqueous ammonia slurry for 3 h reduced the
amount of residual copper to less than 50 ppm.10 However, for
relatively large-scale synthesis (5 kg), after washing for 3 h, the
amount of residual copper was greater than 5000 ppm, and if
washing time was extended, a longer filtration time was
required because grain refining caused clogging of the filter. To
meet large-scale manufacturing requirements, a more effective
scalable method was investigated.
Although it was found that 1b was difficult to solve in various
solvents, our investigation revealed that 1b could be solved in 4
M NaOH aq (5 v/w) and H2O (20 v/w) at 60 °C. Here, it was
assumed that sodium salt 8 was formed, which resulted in the
ring opening of the lactone (Scheme 3).11 Copper salt did not
solve in this solution and was removed by filtration.12 In
addition, it was found that 8 in the filtrate solution was easily
cyclized under acidic conditions (HCl aq) and afforded 1b.
With regard to the suspension at end of the Hurtley reaction, it
was found that adding 4 M NaOH aq and H2O at 60 °C to the
C
DOI: 10.1021/op500334b
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
Scheme 3. Copper removal process by forming soluble 8
reaction suspension was also able to solve all components other
than copper, which could be removed by cellulose powder pad
filtration.13 In 100 kg scale manufacturing, it was confirmed that
residual copper is 4 ppm. Thus, 1b was obtained under acidic
conditions.
The acetyl protecting group could be efficiently removed by
treating 1b with 4-toluenesulfonic acid (PTSA) in EtOH/H2O
to provide the desired 1a·PTSA in 77.8% yield with 99.95%
purity (Scheme 4). Conversion of this reaction was
Byproducts A, B, and C are soluble in acidic filtering
solutions and can be easily removed. The unreacted resorcinol
7 reagent, which caused discoloration, was removed by adding
MeCN (2−2.5 v/w). Residual Na2CO3 was removed under
acidic conditions and emitted as CO2; however, as mentioned
previously, prudent addition is necessary to avoid the foaming
that is associated with large-scale manufacturing processes.
Compared with slurry washing, this purification process is more
favorable because it is not easily influenced by scale.
To establish a robust process, it is important to maintain the
state of the solution. In the reversible reaction of 1b and 8, 1b
was much more dominant than 8. If a sufficiently dissolved
state was not maintained, the loss of yield was observed
resulting from the precipitation of insoluble 1b.
Scheme 4. Deacetylation step by PTSA
quantitative, and there is no impurity; however, the loss of 1a
in the filtrate solution was observed in the crystallization step.
PTSA is used as a reagent in the next step,9 and it was
confirmed that, rather than adding 1 equiv of PTSA, 1a·PTSA
was able to affect the reaction. 1a·PTSA also improved the
filtration property compared with free 1a. In addition, 1a·
PTSA, acted as both a deacetylation reagent and a salt in the
subsequent step. A total of 118.9 kg of 1a·PTSA was prepared
in two steps with 59.5% yield.
Compared with the original medicinal synthesis route, a new
route via the Hurtley reaction reduced the number of steps by
one and improved the total yield from 49.0% to 59.5%. In
addition, reagent cost is reduced from USD3057 to USD196
for manufacturing 1 kg of 1a (Table 5), and the cost of starting
materials is also reduced.4b
For large-scale production, using cellulose powder pad
filtration was expected to take a long time. Therefore, it is
necessary to determine the effect of filtration time. Solubility in
relation to time and assay yield was investigated (Chart 1).
Chart 1. Time dependence of assay yield of 8 in reaction
mixture after adding 4 M NaOH (5 v/w) and H2O (20 v/w)
at 60 °C
CONCLUSION
■
We developed an efficient synthetic method to improve the
Hurtley reaction by using a combination of CuI and Na2CO3
with 5-acetamido-2-bromobenzoic acid. It was revealed that the
complementary effect of CuI, which had high intermolecular
cross coupling activity that inhibited the production of
byproducts because of coupling and the presence of Na2CO3,
led to improved yields of 1b. Removal of residual copper after
the reaction was achieved by filtration with a ring opening
sodium salt 8 that solves in H2O. Sodium salt 8 was easily
closing by acidic conditions and gave 1b with 4 ppm residual
copper. Successful deprotection was achieved by PTSA and
afforded 1a. Because of these improvements, 1a·PTSA
synthesis was achieved in two steps with 59.5% yield. The
cost was drastically decreased by replacing the Pd catalyst with
Cu, using boronic acid rather than resorcinol 7 as the substrate,
and omitting BBr3 in the demethylation.4 In addition, this new
reaction condition was achieved in a multihundred kilogram
scale. Optimization of the Hurtley reaction led to a practical
and scalable synthesis that was more cost-effective and less
labor intensive.
With regard to solubility after the reaction, assay yield was
maintained at more than 70% for 10 h.14 It was observed that
the loss of assay yield was not caused by precipitation of 1b but
was a result of deacetylation of 1a. 1a is a target product in the
deacetylation step, and including 1a is not critical issue. It was
desirable not to change the filtration property in filtering 1b by
including 1a because 1a showed poor filterability.
On the basis of the results to this point, high volume
manufacture of compound 1a was undertaken. In the Hurtley
reaction with 130 kg of 6b, followed by subsequent purification,
it was found that this procedure provided 1b in 76.5% yield
with 93.6% purity (resorcinol 7: 2.4%, 1a: 3.4%, impurity A:
N.D., impurity B: 0.19%, impurity C: 0.24%, others: 0.22%)
with 4 ppm residual copper.
EXPERIMENTAL SECTION
Unless otherwise noted, all materials were obtained from
commercial suppliers and used without further purification. H
■
1
D
DOI: 10.1021/op500334b
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
Table 5. Comparison of cost of reagents in medicinal route and new route
one step
cost for 1a (USD/kg)
two step
cost for 1a (USD/kg)
three step
cost for 1a (USD/kg)
medicinal route
new route
K2CO3
15
BBr3
1756
1756
149
10% Pd/C
H2
210
35
Pd2C12(PPh3)2
subtotal
total
1041
1056
245
3057
resorcinol 7
Na2CO3
CuI
29
1
PTSA·H2O
17
47
subtotal
total
149
196
NMR spectra were recorded on a JEOL GSX400 or on a JEOL
ECP500 spectrometer at ambient temperature. Chemical shifts
are reported as ppm downfield from tetramethylsilane. Data are
reported as follows: chemical shift, multiplicity (s = singlet, d =
doublet, t = triplet, q = quartet, br = broad, m = multiplet),
coupling constants, and number of protons.
HPLC analyses were conducted on a Waters alliance series or
Hitachi LaChrom series. C8 column (3.5 μm; 4.6 mm I.D. ×
150 mm) (X Bridge), isocratic elution with 0.05% TFA in
H2O/0.05% TFA in MeCN (85:15), column temperature 25
°C, 1.0 mL/min, 215 nm, 10 μL injection. Retention times: 6a
(3.2 min), resorcinol 7 (3.4 min), impurity B (4.0 min),
impurity C (4.3 min), 1a (6.0 min), 6b (8.4 min), impurity A
(12.8 min), 1b (18.7 min).
decomposed, 1H NMR (400 MHz, DMSO-d6): δ 2.10 (s, 3H),
6.74 (d, J = 2.4 Hz, 1H), 6.83 (dd, J = 8.8, 2.4 Hz, 1H), 8.00
(dd, J = 8.8, 2.4 Hz, 1H), 8.08 (d, J = 8.8 Hz, 1H), 8.21 (d, J =
8.8 Hz, 1H), 8.50 (d, J = 2.4 Hz, 1H), 10.25 (s, 1H), 10.32 (s,
1H). 13C NMR (100 MHz, DMSO-d6): δ 168.72, 160.56,
159.22, 151.50, 138.74, 130.08, 126.14, 124.26, 122.50, 119.38,
118.13, 113.13, 109.49, 102.91, 24.06. IR (KBr): cm−1 3339,
3252, 3187, 1717, 1665, 1619, 1600, 1548, 1490, 1458, 1408,
1370, 1312, 1274, 1165, 1129, 1026, 896, 836, 777, 726, 632,
550, 467. HRMS: calcd for C15H11NO4 [M+]: 269.0688.
Found: 269.0687.
Manufacturing Synthesis for 8-Amino-3-hydroxy-6H-
benzo-[c]-chromene-6-one·PTSA (1a·PTSA). A mixture of
8-acetamide-3-hydroxy-6H-benzo [c] chromene-6-one 1b (103
kg, 0.383 mol), 4-toluenesulfonic acid monohydrate (218 kg,
1.15 mol), EtOH (1230 kg), and H2O (206 kg) was heated to
75 °C and stirred for 23.5 h. When the reaction was completed
as indicated by HPLC (conv. >99.5%), the reaction mixture
was cooled to −10 °C/h over 7 h, resulting in the precipitation
of the crude product and held at 0 °C to 10 °C for 4 h. The
product was collected by filtration, washed with EtOH (408
kg), and dried in vacuo in an oven to a constant weight at 50
°C to afford compound 1a·PTSA (118.9 kg, yield 77.8%) as a
yellow solid. HPLC purity: 99.95 area %. Mp: 300 °C
General Procedure for the Synthesis of 8-Acetamide-
3-hydroxy-6H-benzo-[c]-chromene-6-one (1b). To a
mixture of 5-acetamido-2-bromobenzoic acid 6b (0.775 mmol
1.0 equiv), resorcinol 7 (2.32 mmol, 3.0 equiv) and H2O (0.640
mL) were added to the bases in Table 3 (2.56 mmol, 3.3
equiv), followed by stirring with heating. After dissolution was
confirmed, the copper catalyst (0.232 mmol, 0.3 equiv) was
added to this solution, and the mixture was stirred for 1−4 h at
50−100 °C. The slurry was stirred at 0−10 °C for 10 min; 6 M
of hydrochloric acid (5 mL) and H2O (5 mL) were added
followed by filtration. The filtered cake was further washed with
H2O (3 mL) and dried in vacuo to afford compound 1b as a
solid.
Manufacturing Synthesis for 8-Acetamide-3-hydroxy-
6H-benzo-[c]-chromene-6-one (1b). To a mixture of H2O
(1300 L), resorcinol 7 (166 kg, 1.51 mol) and 5-acetamide-2-
bromobenzoic acid 6b (130 kg, 0.504 mol) were added 4 M of
a NaOH aqueous solution (145 kg, 0.504 mol). After
dissolution was confirmed, Na2CO3 (123 kg, 1.16 mol) was
added to this solution, followed by stirring at 50 °C. CuI (28.8
kg, 0.151 mol) was added to this slurry, and the mixture was
stirred for 6 h at 50 °C. When reaction was completed as
indicated by HPLC (conv. >98.5%), the reaction mixture was
heated to 60 °C, and 4 M of a NaOH aqueous solution (749
kg) and H2O (2600 L) were added to the reaction solution.
After heating and stirring at 60 °C for 1 h, filtering was
performed using cellulose powder under pressure for 10 h to
remove any insolubles. The reaction mixture was cooled to 30
°C, and acetonitrile (300 kg) and 12 M of hydrochloric acid
(458 kg) were added to maintain the internal temperature at
less than 30 °C, followed by stirring for 3 h. Thereafter, a
precipitated solid was filtered, and the solid was further washed
with H2O (650 L) and dried in vacuo in an oven to a constant
weight at 50 °C to afford compound 1b (103.76 kg, 76.5%) as a
pale gray solid. HPLC purity: 93.6 area %. Mp: 340−353 °C
1
decomposed, H NMR (500 MHz, DMSO-d6) δ: 2.29 (s,
3H), 6.76 (d, J = 2.4 Hz, 1H), 6.85 (dd, J = 8.9, 2.4 Hz, 1H),
6.97 (br s, 3H), 7.11−7.14 (m, 2H), 7.50 (dt, J = 8.0, 1.8 Hz,
2H), 7.57 (dd, J = 8.9, 2.4 Hz, 1H), 7.84 (d, J = 2.4 Hz, 1H),
8.08 (d, J = 8.9 Hz, 1H), 8.22 (d, J = 8.9 Hz, 1H), 10.24 (s,
1H). 13C NMR (100 MHz, DMSO-d6): δ 160.27, 159.71,
151.77, 145.04, 138.14, 135.66, 131.97, 128.27, 128.11, 125.53,
124.64, 123.52, 120.57, 119.95, 113.34, 109.22, 103.03, 20.83.
IR (KBr): cm−1 3327, 2927, 2648, 2016, 1737, 1623, 1547,
1495, 1462, 1369, 1325, 1275, 1209, 1174, 1127, 1011, 901,
845, 817, 778, 726, 689, 570, 538, 460. HRMS: calcd for
C13H9NO3 [M+ − PTSA]: 227.0582. Found: 227.0581.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: kazuhiro.kudou@santen.co.jp. Telephone: +81-749-
48-2924.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Takahiro Honda for his helpful advice on the
manuscript, Ms. Eiko Nishioka for NMR analysis, and Dr.
Takeshi Katsuhira for his helpful advice.
E
DOI: 10.1021/op500334b
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
REFERENCES
■
(1) Schacke, H.; Rehwinkel, H. Curr. Opin. Invest. Drugs 2004, 5,
̈
524−528.
(2) Schottelius, A.; Docke, W. D.; Strehlke, P.; Jaroch, S.; Schmees,
̈
N.; Rehwinkel, H.; Hennekes, H.; Asadullah, K. Proc. Natl. Acad. Sci.
U.S.A. 2004, 101, 227−232.
(3) (a) Masatomo, K.; Yumi, H.; Tomoko, O.; Miwa, T.; Hiroyuki,
A.; Masatsugu, N. J. Ocul. Pharmacol. 2011, 27, 353−360. (b) Matsuda,
M.; Mori, T.; Kawashima, K.; Nagatsuka, M.; Kobayashi, S.;
Yamamoto, M.; Kato, M.; Takai, M.; Oda, T. WO 2007/032556 A1.
(c) Matsuda, M.; Nagatsuka, M.; Mori, T.; Kobayashi, S.; Kato, M.;
Takai, M. WO 2008/059865 A1. (d) Matsuda, M.; Nagatsuka, M.;
Mori, T.; Kobayashi, S. WO 2008/059867 A1.
(4) (a) The cost information of regents: CuI USD 40/kg, resorcinol
7 USD 12/kg, K2CO3 USD 6/kg, Na2CO3 USD 0.5/kg,
Pd2C12(PPh3)2 USD 3300/kg, 10% Pd/C USD 1780/kg, H2 USD
0.029/L, BBr3 USD 260/kg, PTSA·H2O USD 46/kg. The cost
information is referred by actual purchase price. (b) Cost information
of starting materials 2, 3, and 6b is not disclosable, but bromobenzoic
acid 6b and 3 are about the same prices, and boronic acid 2 is much
more expensive than resorcinol 7.
(5) Hurtley, W. R. H. J. Chem. Soc. 1929, 1870−1873.
(6) 8-Nitro-3-hydroxy-6H-benzo [c] chromene-6-one 5 via the
Hurtley reaction was synthesized in 40% yield, see: (a) Krzeszewski,
M.; Vakuliuk, O.; Gryko, D. T. Eur. J. Org. Chem. 2013, 25, 5631−
5644. 8-Amino-3-hydroxy-6H-benzo [c] chromene-6-one derivatives
synthesis via the Hurtley reaction were reported, but not listed isolated
yield, see: (b) Sun, W.; Cama, L. D.; Birzin, E. T.; Warrier, S.; Locco,
L.; Mosley, R.; Hammond, M. L.; Rohrer, S. P. Bioorg. Med. Chem. Lett.
2006, 16, 1468−1472.
(7) All impurity profile was observed by HPLC assay in the filtrate
solution which was filtered off in precipitated solid.
(8) There is no observation of impurities in stirring 60−100 °C for
1−5 h with 6b, resorcinol 7, base, and H2O before adding copper
catalyst.
(9) Residual copper was not able to control N-alkylation in next step,
see: Ohno, A.; Kudou, K.; Katsuhira, T. WO/2014/098046 A1.
(10) To a suspension of crude 1b and H2O (15 v/w), 28% NH3 aq
(5 v/w) was added and stirred for 3 h. Addition of EtOH (3 v/w) and
filtered. The filtered cake was further washed with 2 M HCl (5 v/w)
and H2O (5 v/w) and dried in vacuo to afford 1b as a solid (78%, <50
ppm Cu).
(11) Compound 8 was assumed by 1H NMR in the dryness of
filtered solvent of 1b solution and mass spectrometry in the filtered
1
solution. H NMR (500 MHz, D2O) δ: 2.19 (s, 3H), 6.16−6.18 (m,
2H), 6.91 (d, J = 8.9 Hz, 1H), 7.34 (dd, J = 8.2 Hz, 1H), 7.40 (d, J =
2.1 Hz, 1H), 7.44 (dd, J = 8.2, 2.1 Hz, 1H); MS (ESI): m/z 286 [M −
H]−.
(12) Although qualitative analysis was not studied in copper salt, the
salt color changed to red from light gray white and confirmed a
decrease in weight.
(13) KC FROCK that produced by Nippon Paper Industries, Co.,
Ltd. is available.
(14) Preparation of solubility determination sample: Sampling by
syringe from the suspension and filtered by filter (0.45 μm). To pipet
(1000 μL) immediately, addition of 1 M HCl (2000 μL) and mess up
to 50 mL by MeCN/H2O (2:1).
F
DOI: 10.1021/op500334b
Org. Process Res. Dev. XXXX, XXX, XXX−XXX