Chemoselective Removal of Dimeric 1,3-Diol Impurities Generated from Methyl Grignard Addition onto Esters

Article abstract of DOI:10.1021/op025597c

In two separate cases, we have found that 1,3-diol dimeric impurities were common byproducts generated from the methyl Grignard addition onto esters to form the corresponding tertiary alcohols. These impurities proved difficult to purge by recrystallization as they tended to cocrystallize with the desired product. To avoid the necessity of chromatography, we developed a practical procedure for the chemoselective removal of these 1,3-diol impurities from crude tertiary alcohol by simply performing the recrystallization in the presence of phenylboronic acid. This straightforward method proved equally successful for two cases studied, and the procedure was later demonstrated on a multikilogram scale.

Full text of DOI:10.1021/op025597c

Organic Process Research & Development 2003, 7, 929−932
Chemoselective Removal of Dimeric 1,3-Diol Impurities Generated from Methyl
Grignard Addition onto Esters
John L. Tucker,* Michel Couturier, Kyle R. Leeman, Matthew P. Hinderaker, and Brian M. Andresen
Pfizer Global Research and DeVelopment, Chemical Research and DeVelopment, Eastern Point Road,
Groton, Connecticut 06340, U.S.A.
Table 1. Impurity formation relative to equivalents of
methylmagnesium bromide
Abstract:
In two separate cases, we have found that 1,3-diol dimeric
impurities were common byproducts generated from the methyl
Grignard addition onto esters to form the corresponding
tertiary alcohols. These impurities proved difficult to purge by
recrystallization as they tended to cocrystallize with the desired
product. To avoid the necessity of chromatography, we devel-
oped a practical procedure for the chemoselective removal of
these 1,3-diol impurities from crude tertiary alcohol by simply
performing the recrystallization in the presence of phenyl-
boronic acid. This straightforward method proved equally
successful for two cases studied, and the procedure was later
demonstrated on a multikilogram scale.
entry
equiv MeMgBr
% dimers
1
2
3
2.0
5.0
10.0
15
7
2
Table 2. Attempted purification of 1a
entry
comment
% of 1
% dimers
1
2
3
4
5
6
7
8
crude
IPO
BuOH
EtOAc/hex
THF/hex
MEK
90.70
98.10
96.25
95.00
93.50
95.40
97.17
93.30
88.17
91.60
91.90
1.99
0.88
1.25
1.58
1.67
0.97
0.87
1.53
2.81
2.13
1.94
Introduction
During process development of 1,1 a side reaction was
encountered during methyl Grignard addition onto an ester
(2) that resulted in the formation of dimeric, diol byproducts
(3).2 The impurities proved difficult to purge and remained
in final drug substance at unacceptably high levels (Scheme
1).
ACN
MTBE
pentane
hexanes
heptanes
10
11
12
a
Performed by heating a suspension of 1 in 2 volumes (mL/g 1) solvent.
Sulfonamide 4,3 an advanced intermediate from another
drug candidate, was experiencing identical difficulties as-
sociated with dimeric, diol impurities using similar Grignard
methodology to form a tertiary alcohol from an ethyl ester
(5) (Scheme 2).
These diol byproducts are the result of an alternate
reaction pathway (Scheme 3) during which addition of
methyl Grignard to an ethyl ester can involve enolization of
an intermediate ketone, followed by Aldol-type generation
of dimeric impurities.4
>20% dimeric diol impurities (3).2 The following optimiza-
tions led to an improved reaction profile. Reverse addition
(substrate to an excess of Grignard) at room temperature
resulted in increased Grignard attack, less enolization, and
an improved reaction selectivity for the desired product over
dimer (Table 1).
With the impurity reduced to 2%, selective crystallizations
were explored (Table 2).
Impurities typically crystallized along with the desired
tertiary alcohol. No facile alternative to chromatography
effectively delivered both acceptable purity and yield. Some
crystallization conditions led to a reduction of the impurities,
but not to within drug substance specifications (<0.1%) and
with unacceptably low recovery of product (typically <60%
yield). Other crystallizations delivered higher yield, but
showed little purging ability or even favored enhancement
of the dimeric impurities.
In both cases, purification for removal of these byproducts
required chromatography. Herein, we wish to report a
practical solution for the elimination of 1,3-diol dimeric
impurities generated during methyl Grignard addition to
esters.
Results and Discussion
In the synthesis of 1, addition of 4 equiv of methyl
Grignard to 2 in THF at -10 °C resulted in generation of
Cyclic boronates have been used as directing and protect-
ing groups during synthesis,5 to affect separation of cis from
trans diols,6 and as molecular sensors.7 Although five-
membered cyclic boronates may be formed from cis and trans
* Author for correspondence. E-mail: john.tucker@pfizer.com. Telephone:
858-638-6317. Fax: 858-622-3299.
(1) Marfat, A.; Chambers, R. J.; Watson, J. W.; Cheng, J. B.; Duplantier, A.
J.; Kleinman, E. F. U.S. Patent 6,380,218, B1, 2002.
(2) Characterization via preparative TLC, ES+ LC/MS, and APcI MS.
(3) Dombroski, M. A., Eggler; J. F.; U.S. Patent 6,166,064, 2000.
(4) (a) Batchelor, K. J.; Bowman, R. W.; Davies, R. V.; Hockley, M. H.;
Wilkins, D. J. J. Chem. Res. 1999, 428. (b) Imamoto, T.; Takiyama, N.;
Nakamura, K. Tetrahedron Lett. 1985, 26, 4763.
(5) (a) Evans, D. A.; Polniaszek, R. P. Tetrahedron Lett. 1986, 27, 5683. (b)
Bjorsvik, H.; Priebe, H.; Cervenka, J.; Aabye, A. W.; Gulbrandsen, T.;
Bryde, A. C. Org. Process Res. DeV. 2001, 5, 472.
10.1021/op025597c CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/02/2003
Vol. 7, No. 6, 2003 / Organic Process Research & Development
•
929

Scheme 1
Scheme 2
Scheme 3
Table 4. Results of phenylboronic acid purification as
applied to 4 on pilot plant scale
entry
comment
% of 4
% dimers
1
2
3
crude
THF/IPE
THF/IPE + PhB(OH)2
93.82
96.52
99.06
2.99
2.40
0.07
Table 3. Recrystallization in the presence and absence of
phenylboronic acid
entry
comment
% of 1
% dimers
sociated with dimeric, diol impurities (vide supra). Ap-
proximately 3% of 1,3-diol dimeric impurities (5) remained
in the bulk 4 produced, was carried forward to drug
substance, and could not be removed without chromatogra-
phy. A solvent adjustment was required to accommodate the
characteristics of this new compound (in THF), but the
purification method proved to be equally applicable. De-
rivatization in THF, followed by a displacement into diiso-
propyl ether gave marked improvement in overall purity,
reduced the dimeric impurities, and eliminated the need for
chromatography (Table 4). This procedure was successfully
exemplified on a pilot plant scale to purify 79 kg of 4
(Scheme 4).
1
2
3
crude
toluene
toluene + PhB(OH)2
89.73
96.83
98.39
2.57
1.20
0.03
vicinal diols, angle strain frequently leads to easy hydrolysis.8
1,3-Diols can accommodate both cis and trans isomers in
formation of strainless, six-membered cyclic boronates. Since
boronic acids readily react with 1,3-diols to generate cyclic
boronates, we surmised that it might be possible to selectively
derivatize the dimeric diol impurities from the Grignard
reaction. This could potentially differentiate the impurity to
be removed from the desired tertiary alcohol, allowing for
purification via crystallization. Phenylboronic acid could
provide the requisite boronate, imparting additional lipophi-
licity to the derivatized dimeric diol impurity for greater
solubility in the crystallization solvent.
On the basis of this hypothesis, the crude 1 containing
∼3% dimers was heated to solution at reflux in toluene in
the presence of 10 mol % phenylboronic acid for 1 h. After
cooling to ambient temperature and stirring for 1 h, the
resulting solids were filtered and analyzed. Much to our
delight, only 0.03% dimers remained in the isolated product,
and the dimeric byproducts were found in the mother liquor
as their phenyl boronic esters.9,10 A control experiment (Table
3) was performed in the absence of phenylboronic acid, with
the resultant solid product still containing 1.2% dimeric, diol
byproducts.
Conclusion
We have developed a useful, selective method for removal
of 1,3-diol dimeric impurities encountered during methyl
Grignard attack upon esters. Phenylboronic acid preferentially
forms cyclic boronic esters, which solubilize 1,3-diol dimeric
impurities, allowing for easy purification of tertiary alcohol
products via simple crystallization. In addition, solid-
supported boronic acid reagents11 may conceivably allow for
separation of diol impurities when crystallization of desired
alcohol is not feasible.
Experimental Section
General Procedures. The 1H and 13C NMR spectra were
recorded on a Varian Innova 400 spectrometer. Elemental
analyses were performed by Schwarzkopf Microanalytical
Laboratory, Inc. (Woodside, N.Y.). Melting points were
obtained from a Thomas-Hoover Uni-Melt capillary ap-
paratus and are uncorrected. Samples were analyzed using
reversed-phase HPLC on a Hewlett-Packard (HP1100). The
Sulfonamide 4, an advanced intermediate from another
drug candidate, was experiencing identical difficulties as-
(6) (a) Seymour, E.; Frechet, M. J. Tetrahedron Lett. 1976, 41, 3669. (b)
Tsukagoshi, K.; Kawasaki, R.; Maeda, M.; Takagi, M. C. Chem. Lett. 1994,
681.
(7) Arimori, S.; James, T. D. Tetrahedron Lett. 2002, 43, 507.
(8) Ferrier, R. J. AdV. Carbohydr. Chem. Biochem. 1978, 35, 31-80.
(9) Characterization via preparative TLC and APcI MS.
(10) Boron content of 1 after treatment was measured at 437 ppm, which was
reduced to 20 ppm by a re-slurry in 5 mL of toluene/g of 1 in 89% yield.
(11) (a) Bullen, N. P.; Hodge, P.; Thorpe, F. G. J. Chem. Soc., Perkin Trans. 1
1981, 1863-1867. (b) Liao, Y.; Li, Z. Synth. Commun. 1998, 28, 3539-
3547.
930
•
Vol. 7, No. 6, 2003 / Organic Process Research & Development

Scheme 4
mobile phase buffer consisted of 50 mM KH2PO4 with 25
mM decanesulfonic acid, sodium salt. The buffer was brought
to a pH of 2.5 with phosphoric acid. The mobile phase
consisted of 75% buffer, 15% acetonitrile, and 10% metha-
nol. An Inertsil CN-3 column, 4.6 × 150 × 3 µm3, was used
for the separation. The system flow rate was set at 2.0 mL/
min, and the column temperature was maintained at 40 °C.
The samples were evaluated at a wavelength of 219 nm.
Boron analysis was evaluated using an Agilent ICP-MS
(Inductively Coupled Plasma Mass Spectrometer). Samples
were initially prepared in 5 mL of HNO3 (trace metal grade)
and 5 mL of 30% H2O2. These solutions were then digested
in a microwave labstation provided by Milestone. The
samples were heated using a 50-min temperature gradient
that ranged from ambient to 225 °C. Once fully digested,
the samples were then diluted with 0.5% HNO3. Values were
compared versus a standard Boron curve generated using
Plasma Emission Standards supplied by E. M. Science.
Standard concentrations ranged from 0.01 to 100 ppm. Mass
spectra were obtained on a Waters Micromass ZMD mass
spectrometer using APcI with a 50:50 water/acetonitrile
mobile phase. HPLC mass spectra were obtained on a
Hewlett-Packard HP-1100 series LC/MSD, using a gradient
from 0.01% formic acid in 98:2 water/acetonitrile to 0.005%
formic acid in acetonitrile for mobile phase. A Phenominex
C18-HC column, 30 × 2.00 mm2, 5 µm, was used for
separation.
Crystallization of 2-(4-Fluoro-phenoxy)-N-[4-(1-hy-
droxy-1-methyl-ethyl)-benzyl]-nicotinamide (1). To a 50-
mL round-bottom flask was added 4.97 g (13.07 mmol, 1.0
equiv) of solid 1 (containing 2.6% dimeric diol byproducts)
and 25 mL of toluene. The flask was heated to solution and
held at reflux for 20 min. The solution was allowed to slowly
cool to ambient temperature over 1 h, at which point solids
precipitated. After a 90-min granulation, the solids were
filtered, washed once with 5 mL of toluene, and then placed
in a vacuum oven until dry. Obtained 3.2 g of 1 with 96.8%
overall purity and 1.20% dimeric diol impurities (3) remain-
ing.
held at reflux for 20 min. The solution was allowed to slowly
cool to ambient temperature over 1 h, at which point solids
precipitated. After a 90-min granulation the solids were
filtered, washed once with 5 mL of toluene, and then placed
in a vacuum oven until dry. Obtained 3.13 g of 1 with 98.4%
overall purity and 0.03% dimeric diol impurities (3) remain-
ing.
Re-slurry of 2-(4-Fluoro-phenoxy)-N-[4-(1-hydroxy-1-
methyl-ethyl)-benzyl]-nicotinamide (1) for Removal of
Boron Content. 1 (500 mg, 1.31 mmol, 1.0 equiv containing
437 ppm boron) and toluene (2.5 mL) were added to a round-
bottom flask and stirred at ambient temperature for 36 h.
Toluene (1.5 mL) was added to aid in mobility, and the slurry
was filtered. The solid cake was washed once with 1.2 mL
of toluene. 1 was obtained (446 mg, 89% yield), containing
20 ppm boron content. 1H NMR (400 MHz, CDCl3): δ 8.56
(m, 1H), 8.23 (m, 1H), 8.16 (m, 1H), 7.44 (d, J ) 8.5 Hz,
2H), 7.28 (d, J ) 8.5 Hz, 2H), 7.13-7.03 (m, 5H), 4.65 (d,
J ) 5.5 Hz, 2H), 2.80 (brs, 1H), 1.54 (s, 6H). 13C NMR
(100 MHz, CDCl3): δ 163.74, 161.55, 160.43, 150.14,
148.92, 142.69, 136.58, 127.48, 125.15, 123.73, 123.64,
119.77, 116.91, 116.79, 116.55, 72.41, 43.83, 31.97. Analysis
calculated for C22H21FN2O2: C, 69.46; H, 5.56; N, 7.36.
Found: C, 69.75; H, 5.87; N, 7.45.
2,4-Bis-(4-(2-sulfonamido)furanyl)-2,4-dihydroxypen-
tane (6). Chromatographic isolation provided a small sample
of 2,4-bis-(4-(2-sulfonamido)furanyl)-2,4-dihydroxypentane,
1
5. HNMR (400 MHz, DMSO-d6): δ 7.54 (s, 4), 7.51 (s,
2), 6.83 (s, 2), 5.36 (s, 2), 2.13 (q, 2), 1.35 (s, 6). 13CMR
(400 MHz, DMSO-d6): δ 151.54, 141.23, 136.62, 112.91,
70.27, 53.12, 31.25. LC/MS: m/e 393 (M-1, negative ion).
Purification of 4-(1-Hydroxy-1-methyl-ethyl)-furan-2-
sulfonamide (4). To a 500-mL round-bottom flask was added
50.00 g (243.62 mmol) of 4 (HPLC purity: 93.81% desired
alcohol, 2.99% diol impurity) and 150 mL of THF. The
mixture was stirred for 20 min to ensure maximum dissolu-
tion. The hazy solution was then filtered over 7.5 g of Celite
to remove residual ash, the Celite was rinsed once with 150
mL of THF, and the combined filtrate was added to a 1000
mL of round-bottom flask, followed by 1.53 g (12.17 mmol,
0.05 equiv) of 97% phenylboronic acid. The solution was
distilled until a total volume of 125 mL was reached.
Diisopropyl ether (IPE; 250 mL) was added and THF was
displaced until there was 225 mL of total pot volume. IPE
(225 mL) was added, and displacement continued until there
Purification of 2-(4-Fluoro-phenoxy)-N-[4-(1-hydroxy-
1-methyl-ethyl)-benzyl]-nicotinamide (1). To a 50-mL
round-bottom flask was added 4.97 g (13.07 mmol, 1.0
equiv) of solid 1 (containing 2.6% dimeric diol byproducts),
25 mL of toluene, and 159 mg (1.3 mmol, 10 mol %) of
phenylboronic acid. The flask was heated to solution and
Vol. 7, No. 6, 2003 / Organic Process Research & Development
•
931

1
was 375 mL total pot volume. IPE (125 mL) was added,
and displacement continued until there was 375 mL of total
pot volume. The contents were cooled to 20-25 °C, and
the slurry was allowed to granulate at 20-25 °C for 24 h.
The solids were filtered at 20-25 °C, followed by a wash
with 100 mL of IPE. There was 44.75 g (218.04 mmol, 90%
yield) of eggshell-colored solid after drying. HPLC purity:
99.3% desired alcohol (4), 0.07% diol impurities (5).
Re-slurry of 4-(1-Hydroxy-1-methyl-ethyl)-furan-2-
sulfonamide (4) for Removal of Boron Content. 4 (1.0 g,
4.87 mmol, 1.0 equiv, containing 49 ppm boron) and IPE
(5.0 mL) were added to a round-bottom flask and stirred at
ambient temperature for 36 h. The slurry was filtered. The
solid cake was washed once with 3.0 mL of IPE. This
obtained 959 mg (96% yield) of 4 with undetectable boron
content. H NMR (400 MHz, DMSO-d6): δ 7.67 (s, 3H),
6.95 (s, 1H), 5.08 (brs, 1H), 1.38 (s, 6H). 13C NMR (400
MHz, DMSO-d6): δ 152.33, 140.90, 137.22, 112.83, 67.26,
31.61. Anal. calcd for C22H21FN2O2: C, 40.97; H, 5.40; N
6.82. Found: C, 41.20; H, 5.45; N, 6.79.
Acknowledgment
We thank Dr. Philip Hammen for his insightful comments,
Dr. Jeffrey W. Raggon and Mr. W. Michael Snyder for
exemplifying this purification on pilot-plant scale, and Frank
Urban and V. John Jasys for isolation and characterization
of 6.
Received for review October 8, 2002.
OP025597C
932
•
Vol. 7, No. 6, 2003 / Organic Process Research & Development