Efficient large-scale synthesis of 4-phenyl-3-butyn-2-one, a key intermediate for a novel potent adenosine antagonist

Article abstract of DOI:10.1021/op9700486

Phenylacetylenic Grignard reagent reacts with acetic anhydride under mild conditions to give 4-phenyl-3-butyn-2-one in high yield. This method was applicable to a large-scale synthesis, and optimized reaction conditions have been investigated.

Full text of DOI:10.1021/op9700486

Organic Process Research & Development 1998, 2, 60−62
Technical Notes
Efficient Large-Scale Synthesis of 4-Phenyl-3-butyn-2-one, a Key Intermediate
for a Novel Potent Adenosine Antagonist
Atsuhiko Zanka
Technological DeVelopment Laboratories, Fujisawa Pharmaceutical Co. Ltd., 2-1-6 Kashima,
Yodogawa-ku, Osaka 532, Japan
Scheme 1. Route to FK838 from 4-phenyl-3-butyn-2-one
(1a)
Abstract:
Phenylacetylenic Grignard reagent reacts with acetic anhydride
under mild conditions to give 4-phenyl-3-butyn-2-one in high
yield. This method was applicable to a large-scale synthesis,
and optimized reaction conditions have been investigated.
Results and Discussion
Horiai and co-workers1 have shown that 6-oxo-3-[2-
phenylpyrazolo[1,5-a]pyridin-3-yl]-1(6H)-pyridazinebutyr-
ic acid (FK838), an adenosine A1 receptor antagonist, has
potent diuretic and anti-hypertensive effects which are useful
for the regulation of renal function. We required a large
quantity of FK838 (200 kg) for complete biological evalu-
ation. FK838 is prepared Via the pyrazolo[1,5-a]pyridine
3, the product of a 1,3-dipolar cycloaddition reaction between
ketone 1a and 1-aminopyridinium salt (2, Scheme 1). Whilst
4-phenyl-3-butyn-2-one (1a) can be purchased from the
Aldrich Chemical Co., the required quantity of 1a (>100
kg) was so large that we decided to investigate practical and
inexpensive methods amenable to a large-scale operation.
In this paper we describe a simple, practical, optimized
method for the synthesis of 1a that is readily scaled up to
50-100 kg scale.
The most conceptually simple method for the preparation
of 1a would involve acylation of phenylacetylene; however,
despite various known methods based on this disconnection,
the literature was not helpful in presenting an efficient
procedure. Easily prepared metal acetylides such as Li,2
Mg,3,4 and Na5 suffer from poor yields since the high
reactivity of the organometallic reagent leads to some
undesired products. Furthermore, using heavy metals such
as Cu6 and Zn7 affords good results, but these methods
required tedious procedures and are not favorable from the
Table 1. Effect of acylating reagents
amount of
reagent (equiv)
reagent
yield (%)a
acetic anhydride
N-acetylmorpholine
DMA
2.5
2.5
2.5
2.5
94b
51b
25b
0
AcOMe
AcCl
AcCl
1.25
2.50
0.4c
0.7
a Absolute yield was determined by quantitative HPLC. b Immediately after
the reaction (before addition of MeOH). c 40% of carbinol was obtained.
environmental point of view. Amongst the known methods,
Yamaguchi et al.8 and Brown et al.9 have reported a novel
synthesis of acetylenic ketones Via the reaction of alkynyl-
boron compounds in good yields, but this method requires
low-temperature conditions (-78 °C), and this was not
acceptable to us for a large-scale synthesis.
In order to develop alternative and improved acylating
systems for coupling with alkynylmetal species, we examined
readily available starting materials and inexpensive prepara-
tive methods. Acetyl chloride, N,N-dimethylacetamide,
methyl acetate, and acetic anhydride are inexpensive and
commercially available; thus, we investigated the acylating
ability of these agents. Regarding activation methods for
(1) (a) Horiai, H.; Kohno, Y.; Minoura, H.; Takeda, M.; Nakano, K.; Hanaoka,
K.; Kusunoki, T.; Otsuka, M.; Shimomura, K. Can. J. Physiol. Pharmacol.
1994, 72, P17.3.9. (b) Takeda, M.; Kohno, Y.; Esumi, K.; Horiai, H.;
Ohtsuka, M.; Shimomura, K.; Imai, M. Jpn. J. Pharmacol. 1994, 64, O-376.
(2) Hauptmann, H.; Mader, M. Synthesis 1978, 307.
(3) Gamboni, G.; Theus, V.; Schinz, H. HelV. Chim. Acta 1955, 38, 255.
(4) Kroeger, J. W.; Nieuwland, J. A. J. Am. Chem. Soc. 1936, 58, 1861.
(5) Nightingale, D.; Wadsworth, F. J. Am. Chem. Soc. 1945, 67, 416.
(6) Normant, J. F.; Bourgain, M. Tetrahedron Lett. 1970, 2659.
(7) Verkruijsse, H. D.; Heus-Kloos, Y. A.; Brandsma, L. J. Organomet. Chem.
1988, 338, 289.
(8) Yamaguchi, M.; Shibato, K.; Fujiwara, S.; Hirao, I. Synthesis 1986, 421.
(9) Brown, H. C.; Racherla, U. S.; Singh, S. M. Tetrahedron Lett. 1984, 25,
2411.
60
•
Vol. 2, No. 1, 1998 / Organic Process Research & Development
S1083-6160(97)00048-0 CCC: $15.00 © 1998 American Chemical Society and Royal Society of Chemistry
Published on Web 01/16/1998

Scheme 2. Preparation of carbinol
Table 2. Effect of Ac2O amounta
amount of
Ac2O (equiv)
carbinol
yield (%)
yield (%)b
yield (%)c
10.5
5.0
2.5
2.0
1.25
83
80
94
94
91
complex mixture
1.4
2.2
2.3
2.4
4.3
76
94
a All reactions were performed on a 50 g (0.49 mol) scale. b Absolute yield
was determined by quantitative HPLC immediately after the reaction (before
addition of MeOH). c Absolute yield was determined by quantitative HPLC after
workup (after removal of Mg(OAc)2).
the phenylacetylene, we selected the sodium acetylide and
Grignard reagents, since they are stable and readily prepared
on a large scale.
Table 3. Effect of temperature
amount of
Ac2O (equiv)
carbinol
yield (%)
yield (%)a
Whilst the reaction of the sodium acetylide with acylating
agents suffered from poor yield because of low solubility in
THF, the reaction using the phenylacetylenic Grignard
reagent revealed several characteristic features of this system
(Table 1). Reaction with acetyl chloride afforded mainly
an undesired compound, the carbinol resulting from further
reaction of 1a with the acetylide. Reaction using N,N-
dimethylacetamide required an acidic treatment during
workup to remove the -NMe2 group from the adducts.
However, under these conditions, further reaction of 1a with
hydrogen chloride and/or dimethylamine occurs; thus, the
pH of the reaction mixture must be adjusted quickly to ∼6-
7. Since this could not be done effectively on a large scale,
this method was not pursued. N-Acetylmorpholine gave
some improvement, but the yield was still unsatisfactory from
the point of inexpensive preparation. When an ester was
employed, no product was obtained. As can be seen from
Table 1, the best result was obtained with acetic anhydride.
In previous reports,4 the low yields in this system are believed
to result from reaction of the initially formed ketone with
another equivalent of Grignard reagent to form a carbinol
(Scheme 2).
In order to avoid this second reaction, the Grignard
reagent should be slowly added to a large excess of acetic
anhydride at low temperature. In this way, Kroeger and
Nieuwland4 prepared 1a, but the isolated yield was only
8.3%. On the other hand, from our results, use of a large
excess of acetic anhydride may lead to further reaction of
the product. Indeed, in the case with 10.5 equiv of acetic
anhydride, we did not isolate 1a at all but recovered only
undesired by-products with a small amount of carbinol after
workup of the reaction mixture, in spite of an 83% yield at
the end of the reaction. Examination of the optimum molar
ratio of acetic anhydride afforded the following results. While
use of 2.5 equiv of acetic anhydride afforded undesired
products, 2.0 equiv of of acetic anhydride provided 1a
selectively in 93% yield (88% in the organic layer and 5%
in the aqueous layer). This unexpected difference was due
to the reaction of 1a with acetic acid in the former case.
Therefore, these successful results are attributable to decom-
position of excess acetic anhydride to acetic acid and methyl
acetate by methanol under mild conditions and complete
removal of acetic acid in the form of Mg(OAc)2, followed
by washing with sodium hydroxide solution in water. Thus,
temp (°C)
-5 to +2
-5 to +2
-2 to +4
+3 to +8
94
94
92
88
2.5
2.0
2.5
2.5
3.9
2.4
2.4
5.8
a Absolute yield was determined by quantitative HPLC immediately after the
reaction (before addition of MeOH).
the optimum molar ratio of acetic anhydride proved to be
2.0 equiv (Table 2).
The reaction temperature is also important. According
to the literature,4 very low temperature conditions such as
-25 °C were required in order to avoid the formation of
carbinol. However, our results indicate that this reaction goes
well even at -5 to +2 °C (Table 3); thus, we need no special
equipment for the industrial preparation of 1a.
We have extended this method for the acylation of several
alkynyl derivatives. Based on this optimized method, we
were able to prepare silyl alkynyl ketones as well as alkynyl
ketones. However, the yields for silyl alkynyl ketones were
low. The reasons are not clear, but treatment with water
might decompose the product. The results and isolated yields
1
of these ketones are presented in Table 4. H NMR spectra
of these ketones were in accordance with the assigned
structure and literature data.
Conclusion
In conclusion, we have established a practical and facile
synthesis of 4-phenyl-3-butyn-2-one (1a). The use of an
optimum amount of acetic anhydride as the acyl source is a
significant improvement over the published procedures and
is critical to success. In addition, unlike the conventional
methods, the current procedure offers comparatively mild
reaction conditions and is suitable for large-scale applications.
We have prepared >100 kg of 1a and, hence, >200 kg of
the adenosine A1 receptor antagonist FK838.
Experimental Section
Phenylacetylene is commercially available from the
Wychem Co., was of pure grade, and was distilled before
use to remove a small amount of water.
(10) Kitazume, T.; Ishikawa, N. Chem. Lett. 1980, 1327.
Vol. 2, No. 1, 1998 / Organic Process Research & Development
•
61

Table 4. Alkynyl and silyl alkynyl ketones 1a-g prepared
bp (°C)/ lit. bp (°C)/
yield (%) mmHg mmHg
IR (NaCl) ν
(cm-1
1H NMR
(CDCl3) δ (ppm)
1
R1
R2
)
a
b
C6H5
C6H5
CH3
C2H5
83
70
80/3.5
80/0.5
76-77/0.99 2200, 1670
107/0.68
2200, 1670
2.44 (s, 3H), 7.32-7.57 (m, 5H)
1.22 (t, 3H, J ) 7.34 Hz), 2.70 (q, 2H, J ) 7.40 Hz)
7.36-7.60 (m, 5H)
c
d
e
f
C6H5
tBu
CF3
CH3
CH3
49
61
80
18
26
55/1.0
60/14
70/14
54/14
70/32
86-88/1610 2200, 1704
7.37-7.71 (m, 5H)
1.28 (s, 9H), 2.32 (s, 3H)
1.38-1.64 (m, 4H), 2.30 (s, 3H), 2.37 (t, 3H, J ) 6.91 Hz)
56/159
68/127
50/157
2222, 1680
2212, 1676
nBu
(CH3)3Si CH3
(CH3)3Si C2H5
2152, 1680, 1254 0.23 (s, 9H), 2.34 (s, 3H)
2150, 1684, 1254 1.11 (t, 3H, J ) 7.34 Hz), 2.59 (q, 2H, J ) 7.35 Hz)
g
Preparation of 4-Phenyl-3-butyn-2-one (1a). Ethyl
layer was evaporated under reduced pressure, to the residue
was added CH2Cl2 (397 L), and the resulting solution was
washed with aqueous NaHCO3 (200 L × 1). This organic
layer contained 49.1 kg of 4-phenyl-3-butyn-2-one (1a,
87.8% yield), as determined by quantitative HPLC. Removal
of solvents afforded 1a as an oil which was 93% pure by
HPLC. An analytical sample was obtained by distillation
(78-80 °C/3.5 mmHg; lit.9 76-77 °C/0.9 mmHg). Product
was identified by comparison with an authentic sample
purchased from the Aldrich Chemical Co. by spectroscopic
data (NMR, MS): IR (NaCl, ν [cm-1]) 2200, 1670; 1H NMR
(200 MHz, CDCl3) δ 2.44 (s, 3H), 7.32-7.57 (m, 5H); MS
(EI) m/z 145 (M + H)+.
bromide (2.0 kg, 18.3 mol) and a small amount of iodine
were added to metallic magnesium (11.3 kg, 465 mol) in
dry THF (397 L) and heated to 35 °C. After the mixture
was stirred for 1 h, ethyl bromide (53 kg, 486 mol) was
slowly added at about 35 °C, and the solution was refluxed
for a further 1 h. Control of exotherm in the reaction was
achieved by controlling the addition speed of ethyl bromide
(3 h was required on this scale). Phenylacetylene (39.7 kg,
389 mol) was added to the prepared Grignard reagent at about
35 °C, after which the reaction was continued at the same
temperature until no more ethane was evolved. The prepared
acetylenic Grignard reagent, which was kept at 35 °C to
avoid precipitation, was slowly added to acetic anhydride
(79.3 kg, 777 mol) in THF (198 L) at <4 °C over 1 h. After
additional stirring for 1 h, to this solution was added MeOH
(198 L, 4890 mol). Stirring was continued overnight, and
the precipitate was filtered off and washed with THF (198
L). The filtrate was washed with 1 N NaOH (385 L × 1)
and brine (300 L × 2). The combined aqueous layers
contained 3.0 kg (21 mol) of 1a, and the organic layer
contained 49.6 kg (344 mol) of 1a (HPLC). The organic
Acknowledgment
I thank Dr. David Barrett, Medicinal Chemistry Research
Laboratories, for his interest and advice in this work.
Received for review August 26, 1997.
OP9700486
X Abstract published in AdVance ACS Abstracts, December 1, 1997.
62
•
Vol. 2, No. 1, 1998 / Organic Process Research & Development