Magnesium meerwein-ponndorf-verley-oppenauer reaction. the origin of an impurity in PDA-641 batches

Article abstract of DOI:10.1021/op980078o

An impurity in the initial scale-up batch of PDA-641 was identified and the pathway of its formation was established. The precursor of the impurity was formed by magnesium Meerwein-Ponndorf-Verley-Oppenauer reaction during the Grignard addition step. As a

Full text of DOI:10.1021/op980078o

Organic Process Research & Development 1998, 2, 407−411
Magnesium Meerwein-Ponndorf-Verley-Oppenauer Reaction. The Origin of
an Impurity in PDA-641 Batches
Bogdan K. Wilk,*,† Jean L. Helom,† and Clifford W. Coughlin‡
Chemical DeVelopment, Wyeth-Ayerst Research, 401 North Middletown Road, Pearl RiVer, New York 10965, and
64 Maple Street, Rouses Point, New York 12979
Chart 1
Abstract:
An impurity in the initial scale-up batch of PDA-641 was
identified and the pathway of its formation was established.
The precursor of the impurity was formed by magnesium
Meerwein-Ponndorf-Verley-Oppenauer reaction during the
Grignard addition step. As a result of the investigation, the
level of the impurity could be reduced 6-fold in the subsequent
batch.
Introduction
Inhibitors of phosphodiesterase IV isoenzyme (PDE IV)
have therapeutic advantage in the management of asthma.1
The discovery at Wyeth-Ayerst Research of a preferential
PDE-IV inhibitor, 3′-(cyclopentyloxy)-4′-methoxyacetophe-
none (E)-O-carbamoyloxime (PDA-641; 1, Chart 1), gener-
ated a need for multikilogram quantities of the drug
substance.
The synthetic pathway to 1 included MeMgCl addition
to 3-(cyclopentyloxy)-4-methoxybenzaldehyde (2, Chart 1),2b
followed by hypochlorite oxidation, oximation, and carbam-
oylation (Scheme 1). The initial scale-up batch of PDA-
641 contained an unknown impurity at a level of 0.50%sa
borderline value for the purity requirements.3 The impurity
was difficult to remove by recrystallization. Therefore, it
was necessary to identify the impurity, its source, and the
way of suppressing its formation.
Results and Discussion
The LC-MS data suggested that the impurity might be
the carbamate of the aldehyde oxime, 3 (Chart 1). Its identity
was confirmed by chromatographic and spectroscopic com-
parison to an authentic sample of 3. It was presumed that
the precursor to impurity 3 was the analogous oxime 5 (Chart
1). In fact, it was present as the largest single impurity in
all the oxime 6 (Chart 1) batches. An authentic sample of
5 was prepared from 2 and characterized spectroscopically.
Once generated, 5 was converted to the carbamate 3 in the
carbamoylation step.
Scheme 1
Initially, it was thought that, prior to oximation, the
presence of aldehyde 2 in ketone 4 (Chart 1)2b solutions was
the result of incomplete MeMgCl addition. Although the
residual aldehyde route cannot be ruled out, there is another
pathway by which the aldehyde is generated (Scheme 2).
The addition of MeMgCl to 2 is the principal reaction
pathway. The formed chloromagnesium sec-alcoholate 7-Mg
reacts with aldehyde 2 to give the chloromagnesium alco-
holate of primary benzylic alcohol 8-Mg and ketone 4.
† Wyeth-Ayerst Research, Pearl River, NY.
‡ Wyeth-Ayerst Research, Rouses Point, NY.
(1) Cavalla, D.; Frith, R. Curr. Med. Chem. 1995, 2, 561.
(2) (a) Heaslip, R. J.; Lombardo, L. J.; Golankiewicz, J. M.; Ilsemann, B. A.;
Evans, D. Y.; Sickels, B. D.; Mudrick, J. K.; Bagli, J. F.; Weichman, B. M.
J. Pharmacol. Exp. Ther. 1994, 268, 888. (b) Lombardo, L. J. Eur. Pat.
Appl. 470 805 A1, 1992.
(3) Wyeth-Ayerst Research internal specification.
10.1021/op980078o CCC: $15.00 © 1998 American Chemical Society and Royal Society of Chemistry
Published on Web 10/29/1998
Vol. 2, No. 6, 1998 / Organic Process Research & Development
•
407

Scheme 2
Based on LC-MS mass fragmentation pattern, primary
alcohol 8, ketone 4, and tertiary alcohol 9 were present in
the batch. Synthesis of authentic samples of 8 and 9
confirmed both structures. Thus formed, primary alcohol 8
was carried over to the oxidation step, where it underwent
oxidation to aldehyde 2 but not to the acid. Under hy-
pochlorite phase-transfer conditions, primary benzylic alco-
hols undergo oxidation to aldehydes.4 Finally, 2 was
converted to 5 and then to 3 in the subsequent steps. To
suppress formation of 3, it was imperative to minimize
generation of primary alcohol 8 at the Grignard stage.
The reaction described in Scheme 2 resembles the
Cannizzaro reaction,5 a base-catalyzed dismutation of an
aromatic aldehyde.6 It was our initial preference to use the
term “magnesium-Cannizzaro reaction” as a distinction from
Meerwein-Ponndorf-Verley-Oppenauer (MPVO) reac-
tions.7 These transformations involve hydride-transfer mech-
anism. However, MPVO reactions usually take place with
group III and IV elements. Magnesium can be considered
as a special case of MPVO reactions. Indeed, magnesium
ethoxide was used by Meerwein and Schmidt to reduce
benzaldehyde.8 Marshall was the first to observe and
recognize an excess of benzaldehyde as an oxidant in
reactions with the Grignard reagents.9 He also pointed out
the similarities to the Cannizzaro reaction and proposed the
mechanism. Later, Shankland and Gomberg10 and Meisen-
heimer11 supplied further evidence in favor of Marshall’s
proposal. More recently, an example of magnesium-
Oppenauer oxidation was described by Byrne and Karras.12
The reaction was also used as a method of ketone prepara-
Scheme 3
tion.13 We decided to use the term “Mg-MPVO reaction”
because of the mechanistic similarities.14
The MPVO transformations are equilibrium reactions
which show a strong preference for the formation of primary
alcoholates and ketones in equilibria involving aldehydes and
secondary alcoholates.15 The initial step involves complex-
ation of the carbonyl group with the metal atom.16 The
activated aldehyde then accepts the hydride from the
magnesium alcoholate (Scheme 3).17
Alkoxy groups on the aromatic ring of the hydride donor
7-Mg facilitate hydride transfer from the benzylic position.
Moreover, para-substitution considerably increases the reac-
tion rate and equilibrium constant.18 The aromatic aldehyde
2, an acceptor, has a low reduction potential, placing it among
good oxidants.19
It was necessary to explore some details of the Mg-MPVO
reaction of 2 with 7-Mg and MeMgCl. Product distribution
data of the addition of MeMgCl to aldehyde 2 in THF
showed that higher temperatures favored the formation of
the primary alcohol 8 (1.2 and 9.0% at 0 and 65 °C,
respectively). Lowering the temperature to -5 °C sup-
pressed the formation of 8 to 0.5%. Addition of MeMgCl
(13) Bradsher, C. K. J. Am. Chem. Soc. 1944, 66, 45. Bachmann, W. E. J. Am.
Chem. Soc. 1934, 56, 1363.
(4) Skarzewski, J.; Siedlecka, R. Org. Prep. Proc. Int. 1992, 24, 625. The
aldehyde 2 was subjected to the reaction conditions and was recovered
thereafter.
(5) Gokel, G. W.; Gerdes, H. M.; Rebert, N. W. Tetrahedron Lett. 1976, 653.
(6) An analogous product arising from the Tishchenko reaction was not
observed.
(7) de Graauw, C. F.; Peters, J. A.; van Bekkum, H.; Huskens, J. Synthesis
1994, 1007.
(8) Meerwein, H.; Schmidt, R. Justus Liebigs Ann. Chem. 1925, 444, 221.
(9) (a) Marshall, J. J. Chem. Soc. 1914, 105, 527; (b) 1915, 107, 509; (c) 1925,
127, 2184.
(14) The reaction could have been named “the Marshall reaction”. Nomenclature
choice between “Meerwein-Ponndorf-Verley reduction” and “Oppenauer
oxidation” depends upon the transformation of the main substrateswhether
it is reduced or oxidized.
(15) Morrison, J. D.; Ridgway, R. W. J. Org. Chem. 1974, 39, 3107.
(16) Dradi, E.; Pochini, A.; Ungaro, R. Gazz. Chim. Ital. 1980, 110, 353.
(17) (a) Kharasch, M. S.; Reinmuth, O. Grignard Reactions of Nonmetallic
Substances; Pentice-Hall, Inc.: New York, 1954; pp 158-160. (b) Ref 17a,
pp 166-176. (c) Moulton, W. N.; van Atta, R. E.; Ruch, R. R. J. Org.
Chem. 1961, 26, 290.
(18) Knauer, B.; Krohn, K. Liebigs Ann. 1995, 677.
(10) Shankland, R. V.; Gomberg, M. J. Am. Chem. Soc. 1930, 52, 4973.
(11) Meisenheimer, J. Justus Liebigs Ann. Chem. 1926, 446, 76.
(12) Byrne, B.; Karras, M. Tetrahedron Lett. 1987, 28, 769.
(19) See ref 7. Adkins, H.; Cox, F. W. J. Am. Chem. Soc. 1938, 60, 1151. Adkins,
H.; Elofson, R. M.; Rossow, A. G.; Robinson, C. C. J. Am. Chem. Soc.
1949, 71, 3622.
408
•
Vol. 2, No. 6, 1998 / Organic Process Research & Development

Table 1. Product distribution (HPLC area, %) in MeMgCl
and 7-Mg reactions with 2
order of
addition
reaction temp
time (h) (°C)
8
7
9
2 + 4a
1
2
3
4
5
6
7
8
9
MeMgCl to 2
MeMgCl to 2
1
1
-60
-10
0.3 74.4
0.5 96.8
20 16.2 52.4 6.5 24.1
0
0
24.5
0.7
MeMgCl to 2b 24
MeMgCl to 2
MeMgCl to 2
MeMgCl to 2
2 to MeMgCl
2 to MeMgCl
2 to 7-Mg
23c
2
0
0
0
0
20
0
32.7 49.4 2.2 13.2
1.1 97.1 0.2
0.4 96.9 0.2
1.0
0.7
8.8
0.75d
1
1
3.2 87.6 0
5.3 80.1 0.2 14.2e
3.75f
3.5
29.8 34.4
0
0
34.2
26.5
10 2 to 7-Mg
25 37.3 34.5
a Both 2 and 4 were inseparable under the HPLC conditions. b In Bu2O/tert-
BuOMe mixture; nonhomogeneous. c Total addition time. d MeMgCl added in
2 min. e Consisted of ∼8.5% of 2 and 5.7% of 4 (peaks not fully resolved).
f The reaction mixture was kept at 4 °C for 2 days.
at -60 °C, followed by quench with methanol at this
temperature and then with water, gave only 0.4% of primary
alcohol 8, but it also left unreacted aldehyde (entry 1, Table
1). Tetrahydrofuran is a Lewis base20 which decreases the
electronegativity of the metal, lowering the reaction rate of
hydride transfer.12 In less complexing ethers such as Bu2O
and t-BuOMe, more 8 was formed (entry 3, Table 1).
Interestingly, reverse addition, 2 to MeMgCl, gave more 8
than the normal addition, MeMgCl to 2 (entries 7 and 8,
Table 1). Apparently, this was due to the higher viscosity
of the reaction medium, especially at lower temperatures.
This may also indicate relatively slow reaction rates for the
addition of MeMgCl to 2. It was observed that the product
distribution depended upon the time of MeMgCl addition.
The kinetics of the addition were investigated, revealing
substantial formation of primary alcohol 8 when MeMgCl
was added at very slow rates (entries 4-6, Table 1). Faster
additions gave lesser amounts of 8. A large excess of 2 at
the beginning of MeMgCl addition shifts the Mg-MPVO
equilibrium to the right side (Scheme 2). This shift is
enhanced by removal of 4 due to its further reaction with
MeMgCl to form 9-Mg.21
Figure 1. Formation of 8 in the reaction of 7-Mg with 2 in
THF at room temperature (HPLC area, %).
Figure 2. Batch (0) and jacket (]) temperature vs addition
time of MeMgCl to 2 (batch A).
The Mg-MPVO reaction was modeled using magnesium
sec-alcoholate 7-Mg and aldehyde 2 (entries 9 and 10, Table
1). Slow addition of the alcoholate at both 0 and 25 °C
gave 30-37% of the primary alcohol 8. During the reaction
of equimolar amounts of 7-Mg and 2 at room temperature
in THF, formation of the primary alcohol 8 was initially
relatively fast22 and then slowed substantially, reaching the
equilibrium (Figure 1).
the temperature was kept at ca. -10 °C. Although the total
addition time of MeMgCl was shorter in the first batch
(Figures 2 and 3), the reagent was added at a higher rate
toward the end of the step in the second batch. Comparison
of HPLC runs after completion of the Grignard step showed
a decrease in the amount of primary alcohol 8 (Table 2).
The data collected during the first plant batch in conjunc-
tion with laboratory results lead to better control of the
second batch. In the first pilot plant batch (A), MeMgCl
was added at about -6 °C, whereas in the second batch (B),
Conclusions
To minimize formation of impurity 3, the primary alcohol
8 that was formed by Mg-MPVO reaction pathway had to
be reduced. The only variables to control the reaction were
temperature and addition time. Chemical properties of the
substrates favored Mg-MPVO reaction in competition with
the Grignard addition. The solvent (THF) had the opposite
effect. The decrease in the amount of 8 was achieved by
reasonably fast (2-3 h) addition of MeMgCl solution in THF
at low temperature (<-6 °C). As result, the second
(20) Isaacs, N. S. Physical Organic Chemistry; John Wiley & Sons: New York,
1987; pp 196-198.
(21) The presence of ketone 4 in the reaction mixture and minimal formation of
tertiary alcohol 9 can be explained by a reaction of the former with the
base (chloromagnesium sec-alcoholate 7-Mg or MeMgCl) to form an enolate.
Cf. refs 11, 17b. In fact, a slow, dropwise addition of MeMgCl into 4 in
THF produced extensive frothing, and 4 was recovered upon quenching.
Ketone 4 can also be in a complex prior to hydrolysis. See ref 9b.
(22) These reactions appear to be rather rapid. Cf. ref 10.
Vol. 2, No. 6, 1998 / Organic Process Research & Development
•
409

recrystallized from THF-heptane to give a light yellow solid
(1.5 g, 27% yield; purity 96.5% by HPLC area).
1H NMR (CDCl3/DMSO-d6): δ 8.25 (s, 1H), 7.26 (s, 1H),
7.15 (d, J ) 8.2 Hz, 1H), 6.89 (d, J ) 8.3 Hz, 1H), 6.35 (br
s, 2H), 4.84 (m, 1H), 3.89 (s, 3H), 2.00-1.81 (m, 6H), 1.62
(m, 2H). 13C NMR (CDCl3/DMSO-d6): δ 142.6, 139.5,
139.3, 134.1, 109.2, 108.8, 99.0, 97.7, 66.6, 42.1, 18.8, 10.2.
UV: λmax 210 nm (ꢀ 67.5). IR (KBr): 3451, 3411 (NH2),
1724 (CdO), 1261, 1016 (C-O), 954 (N-O) cm-1. EI-
MS: m/z (relative intensity) 278 (M+), 277 (M+ - H), 209,
167 (100), 149. Anal. Calcd for C14H18N2O4: C, 60.42; H,
6.53; N, 10.06. Found: C, 60.32; H, 6.64; N, 10.19.
Aldehyde oxime 5 was prepared according to the known
procedure.2b Thus, the aldehyde 2 (8.8 g; 40 mmol) gave
9.9 g of a light yellow oil which solidified on standing to a
waxy, cream-colored solid. The material was a mixture of
the E and Z isomers. The two isomers separated on TLC
(Rf Z ) 0.22, E ) 0.36), whereas HPLC scan showed only
one peak with 99.7 area %, most likely the E isomer.
Milligram quantities of each isomer were isolated by Prep
LC (Waters LC 2000; PrepPak Porasil 15-20 µm, 125 Å;
hexanes-EtOAc (5:1 v/v)) and the structures confirmed by
NMR and MS. A solution of the Z isomer in chloroform-d
was rescanned after 18 h, showing almost complete conver-
sion to the E isomer.
Figure 3. Batch (0) and jacket (]) temperature vs addition
time of MeMgCl to 2 (batch B).
Table 2. Product distribution (HPLC area, %) in batches A
and B after MeMgCl addition
batch
8
7
9
2 + 4a
A
B
2.7
1.4
95.8
98.0
0
0.4
1.5
0.2
1H NMR: δ (E) 8.94 (br s, 1H), 8.08 (s, 1H), 7.21 (s,
1H), 7.02 (d, J ) 8.2 Hz, 1H), 6.84 (d, J ) 8.3 Hz, 1H),
4.81 (m, 1H), 3.87 (s, 3H), 2.06-1.78 (m, 6H), 1.58 (m,
2H); (Z) 7.72 (s, 1H), 7.41 (dd, J ) 8.4, 2.0 Hz, 1H), 7.27
(s, 1H), 6.89 (d, J ) 8.5 Hz, 1H), 4.81 (m, 1H), 3.89 (s,
3H), 1.97-1.83 (m, 6H), 1.61 (m, 2H). 13C NMR: δ (E)
148.3, 146.8, 144.6, 121.4, 118.0, 108.0, 77.1, 52.6, 29.3,
20.7. UV: λmax 217 nm (ꢀ 91.4). IR (KBr): 3475 (OH),
1625 (CdN), 948 (N-O) cm-1. EI-MS: m/z (relative
intensity) 235 (M+), 167 (100), 152, 124. Anal. Calcd for
C13H17NO3: C, 66.36; H, 7.28; N, 5.95. Found: C, 66.64;
H, 7.45; N, 6.01. In support of the structure, oxime 5 was
converted to its acetate.24 Anal. Calcd for C15H19NO4: C,
64.96; H, 6.92; N, 5.05. Found: C, 65.02; H, 7.11; N, 5.09.
Secondary alcohol 7 was prepared according to the
published procedure.2b
1H NMR: δ 6.93 (d, J ) 1.8 Hz, 1H), 6.88 (dd, J ) 8.2,
1.8 Hz, 1H), 6.83 (d, J ) 8.2 Hz, 1H), 4.80 (m, 2H), 3.84
(s, 3H), 1.99-1.80 (m, 6H), 1.61 (m, 2H), 1.48 (d, J ) 6.4
Hz, 3H). 13C NMR: δ 149.2, 147.6, 138.5, 117.5, 112.3,
111.7, 80.3, 70.0, 56.0, 32.7, 25.0, 24.0.
Primary Alcohol 8. A solution of sodium borohydride
(0.1 g; 3 mmol) in 20% sodium hydroxide (10 mL) was
added dropwise into a stirred solution of 2 (1.1 g; 5 mmol)
in 2-propanol (5 mL) in an ice bath. After the reaction was
completed (TLC), the mixture was acidified with 2 M HCl
and extracted with ether. The organic phase was washed
with water and dried over MgSO4. Filtration followed by
evaporation gave 8 as an oil (0.9 g; 81% yield). An
analytical sample was obtained using flash column chroma-
tography (purity 98.1% by HPLC area).
a Both 2 and 4 were inseparable under the HPLC conditions.
improved batch contained only 0.09% of impurity 3, as
compared to 0.50% in the first batch.
Experimental Section
All the reagents and solvents were used as purchased,
except where indicated. 1H NMR spectra were determined
on a Bruker Avance DPX 300 (300.13 MHz) spectrometer.
Chemical shifts are reported in parts per million (ppm)
relative to residual CHCl3 (7.26 ppm). 13C NMR spectra
were recorded at 75.47 MHz. Carbon chemical shifts are
reported in parts per million (ppm) relative to residual CHCl3
(77.09 ppm). IR spectra were obtained on a Mattson RS-1
FT-IR spectrometer and are uncalibrated. UV spectra were
obtained on a Hewlett-Packard HP-8450A spectrophotometer
as solutions in acetonitrile. Electron ionization mass (EI-
MS) spectra were obtained on a Finnigan MAT 90 spec-
trometer. Flash column chromatography was carried out on
J.T. Baker 40 µm silica gel using hexanes-EtOAc (2:1 v/v).
HPLC scans were obtained on a Hitachi D-6000 instrument
with a Partisil 5 µm, ODS-3, 4.6- × 250-mm column
(Whatman); phosphate buffer pH 3.5-acetonitrile (3:2 v/v);
flow 1.0 mL/min; wavelength 226 nm.
Aldehyde Oxime Carbamate 3. Purified 5 (4.4 g)
underwent carbamoylation according to the literature pro-
cedure.23 The aqueous solution was decanted, and the
remaining oil was slurried in a toluene-heptane mixture.
The resulting yellow solid was washed with heptane and
dried in vacuo to give 1.7 g of a crude product. It was
(23) Loev, B.; Kormendy, M. F. J. Org. Chem. 1963, 28, 3421.
(24) Barbry, D.; Champagne, P. Synth. Commun. 1995, 25, 3503.
410
•
Vol. 2, No. 6, 1998 / Organic Process Research & Development

1H NMR: δ 6.92 (s, 1H), 6.88 (d, J ) 8.2 Hz, 1H), 6.84
(d, J ) 8.1 Hz, 1H), 4.80 (m, 1H), 4.61 (s, 2H), 3.84 (s,
3H), 2.13-1.78 (m, 6H), 1.61 (s, 2H). 13C NMR: δ 149.4,
147.6, 133.6, 119.3, 114.0, 111.7, 80.3, 65.0, 56.0, 32.7, 24.0.
IR (neat): 3391 (OH) cm-1. EI-MS: m/z (relative intensity)
222 (M+), 154 (M+ - C5H8; 100). Anal. Calcd for
C13H18O3: C, 70.24; H, 8.16. Found: C, 69.99; H, 8.12.
Tertiary Alcohol 9. To a solution of 4 (0.2 g; 1.0 mmol)
in THF (4.5 mL) was added a solution of methylmagnesium
chloride in THF (3 M; 3 mL; 9 mmol) in one portion from
a syringe. The reaction was exothermic. It was allowed to
stir at room temperature for 30 min (a single spot by TLC),
quenched into water (30 mL) in a 250-mL Erlenmeyer flask,
and extracted with ethyl acetate (70 mL). The organic phase
was washed with water, separated, and dried over MgSO4.
Filtration followed by evaporation gave an oil which was
passed through a small silica gel column. Evaporation of
the eluent, followed by flash column chromatography, gave
9 (0.2 g; 83% yield; purity 99.0% by HPLC area).
24.0. IR (neat): 3427 (OH) cm-1. EI-MS: (relative
intensity) m/z: 250 (M+), 232 (M+ - H2O), 182 (M+
-
C5H8), 167 (M+ - C5H8 - CH3; 100), 164 (M+ - H2O -
C5H8). Anal. Calcd for C15H22O3: C, 71.97; H 8.86.
Found: C, 71.91; H 8.86.
Reactions of MeMgCl and 7-Mg with 2. The experi-
ments were run on a 7-10 mmol scale. The analytical
aliquots (10 µL) were quenched into the HPLC mobile phase
(2 mL), homogenized with Vortex, filtered through Acrodisc
CR PTFE (0.45 µm), and analyzed on a Hitachi D-6000
instrument.
Acknowledgment
We thank Wyeth-Ayerst Research Analytical Department
for support, especially Mr. I. Greene and Dr. G. Goyal. We
also thank Ms. N. Mwisiya for laboratory support. A diligent
and dedicated effort of the pilot plant operators, Messrs. M.
Pivetta, G. Manor, and M. Juneau, is strongly appreciated.
1H NMR: δ 7.08 (s, 1H), 6.98 (d, J ) 8.4 Hz, 1H), 6.84
(d, J ) 8.3 Hz, 1H), 4.80 (m, 1H), 3.84 (s, 3H), 2.04-1.78
(m, 6H), 1.62 (m, 2H), 1.57 (s, 6H). 13C NMR: δ 148.8,
147.3, 141.9, 116.4, 112.2, 111.5, 80.5, 72.2, 56.1, 32.8, 31.7,
Received for review August 18, 1997.
OP980078O
Vol. 2, No. 6, 1998 / Organic Process Research & Development
•
411