Regioselective addition of mesitol to a 2,4-dichloropyridine

Article abstract of DOI:10.1021/op800004q

The regioseleethity of the addition of 2,4,6-trimethylphenol to 2,4-dichloro-3,6-dimethylpyridine can be controlled by the proper choice of catalyst and solvent. The use of catalytic eopper(I) salts and pyridine as solvent results in exclusive addition at

Full text of DOI:10.1021/op800004q

Organic Process Research & Development 2008, 12, 411–413
Regioselective Addition of Mesitol to a 2,4-Dichloropyridine
Sally Gut Ruggeri,* Brian C. Vanderplas, Bruce G. Anderson,† Ralph Breitenbach, Frank J. Urban, A. Morgan Stewart, III, and
Gregory R. Young
Chemical Research and DeVelopment, Pfizer Global Research and DeVelopment, Pfizer Inc, Eastern Point Road,
Groton, Connecticut 06340, U.S.A.
Scheme 2. Addition of 3-pentanol to
2-4-dichloro-3,6-dimethylpyridine
Abstract:
The regioselectivity of the addition of 2,4,6-trimethylphenol to 2,4-
dichloro-3,6-dimethylpyridine can be controlled by the proper
choice of catalyst and solvent. The use of catalytic copper(I) salts
and pyridine as solvent results in exclusive addition at C-2. In their
absence, a mixture of regioisomers is obtained in which addition
at C-4 is dominant.
Table 1. Addition of 2,4,6-trimethylphenol to
Introduction
2-4-dichloro-3,6-dimethylpyridinea
Corticotropin-releasing factor (CRF) antagonists have been
suggested for use in many therapeutic areas, including depres-
sion, anxiety and stress-related diseases.1 4-(1-Ethylpropoxy)-
3,6-dimethyl-2-(2,4,6-trimethylphenoxy)pyridine (1) is a potent
CRF antagonist discovered at Pfizer.2 The original synthetic
route is shown in Scheme 1. 2,4-Dihydroxy-3,6-dimethylpyri-
dine3 (2) was converted to its dichloro analogue4 (3) by reaction
with POCl3/diethylaniline. Aryl ether formation was achieved
by addition of 2,4,6-trimethylphenol (mesitol) sodium salt in
DMSO, but the undesired C-4 regioisomer was the major
product of the reaction. On laboratory scale, the regioisomers
could be separated by column chromatography, and the minor
isomer (4) was reacted with the sodium or potassium salt of
3-pentanol in DMSO to give the desired compound. The low
entry
base
additive
solvent
conversion
4:5b
1:3
100:0
0:100
1:4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
NaH
NaH
-
DMSO
toluene
DMF
DMAc
DMSO
DMSO
DMSO
THF
DMSO
DMSO
diglyme
DMSO/py
DMF/py
pyridine
pyridine
pyridine
>99%
trace
20%
93%
91%
85%
0%
-
LiOH
KOt-Bu
Cs2CO3
n-Bu4NOHc
NaH
NaH
NaH
NaH
n-BuLi
NaH
-
-
-
1:4
1:4
-
Ag2CO3
BF3•OEt2
CuBr
CuBr2
CuCl
CuBr
Cu bronze
-
-
0%
-
50-67%
33%
trace
33%
30%
>95%
0%
2:1
2:3
100:0
8:1
1:3
4:1
KOt-Bu
KOt-Bu
-
CuBr
-
KOt-Bu
Cu(I)d
100%
100:0e
a Typical reaction conditions: Equimolar amounts of 3 and the phenol were
mixed with 1.1 equiv of the base in the solvent and heated to reflux if the bp of
the solvent was <120 °C or 120-150 °C if high-boiling solvents were employed.
The additives were initially screened using 1.0 equiv, and were added prior to
heating. Conversions shown are generally after 18-24 h. b Ratios determined by
GC/MS. c n-Tetrabutylammonium salt of phenol was preformed separately and
isolated prior to reaction. d CuCl, CuBr, and CuI were used interchangeably. e No 5
was observed; however, the product of double displacement was seen.
Scheme 1. Original Synthesis of 1
overall yield and chromatographic separation were not optimal
for the preparation of larger quantities needed for clinical
development. This prompted an investigation into the control
of regioselectivity of the chloride displacement by aryloxides
and alkoxides.
A seemingly trivial improvement to the regioselectivity issue
would be to reverse the order of oxygen nucleophile additions
and add 3-pentoxide first to the dichloropyridine. Unfortunately,
the sodium salt of 3-pentanol added preferentially at C-2,
affording 7 as the major isomer in a 6:1 ratio (Scheme 2).
At this point, a more thorough investigation examining the
effects of counterion and solvent in the addition of mesitol to
3 was undertaken. The highlights of that investigation are shown
in Table 1. In most cases, simply altering the counterion,
whether an alkali metal or a quaternary ammonium salt, had
little of the desired effect on the regioselectivity, and polar, high-
boiling solvents were required to get reasonable conversion
* Author to whom correspondence may be sent. E-mail: sally.gut@pfizer.com
† Current address unknown.
(1) Chen, Y. L.; Mansbach, R. S.; Winter, S. M.; Brooks, E.; Collins, J.;
Corman, M. L.; Dunaiskis, A. R.; Faraci, W. S.; Gallaschun, R. J.;
Schmidt, A.; Schultz, D. W. J. Med. Chem. 1997, 40, 1749.
(2) Chen, Y. L. Corticotropin Releasing Factor Antagonists. U.S. Patent
5,962,479, October 5, 1999.
(3) Mittelbach, M.; Schmidt, H.-W.; Uray, G.; Junek, H.; Lamm, B.;
Ankner, K.; Brandstrom, A.; Simonsson, R. Acta Chem. Scand. B 1988,
524.
(4) Prelog, V.; Szpilfogel, S. HelV. Chim. Acta 1942, 25, 1306.
10.1021/op800004q CCC: $40.75
Published on Web 04/05/2008
2008 American Chemical Society
Vol. 12, No. 3, 2008 / Organic Process Research & Development
•
411

(entries 1, 3-6); nonpolar solvents seemed to favor the desired
isomer, but the reaction was extremely slow. The major product
where good conversion was achieved was the undesired C-4
isomer, 5. However, in our investigation of additive counterions,
we noted that addition of copper salts (entry 9) to the standard
reaction had a beneficial effect on the regioselectivity and
inverted the ratio in favor of attack at C-2 to give 4 as the major
isomer. In another experiment, adding a small amount of
pyridine to the reaction in order to solubilize the copper salts
further increased the ratio of products in favor of 4 (entry 12).
In an attempt to separate counterion from solvent effects, we
ran the reaction with potassium tert-butoxide in pyridine in the
absence of copper catalyst (entry 14). Desired isomer 4 was
still the major one, but to a lesser extent. When pyridine was
used as solvent with potassium tert-butoxide and a catalytic
amount of a copper salt, none of the undesired regioisomer could
be detected. The only other product of the reaction arose from
displacement of both chlorides by the phenol. The amount of
the bis-adduct formed could be minimized by adjusting the
stoichiometry of mesitol and base relative to 3. Under the
optimized conditions (1.1 equiv of each), <5% of this bis-adduct
was formed, and was largely purged during isolation.
A rate enhancement was also observed. In all the other
reactions shown, the ether formation was sluggish, and often
failed to go to completion even after 24 h. Under the preferred
conditions (25 mol % catalyst), the reaction was complete within
2 h. As little as 5 mol % was sufficient to drive the reaction to
completion, but a tailing in conversion was observed, and the
reaction required overnight reflux to consume the starting
materials. All of the copper(I) halide sources tried gave
equivalent results. Copper(0) or -(II) sources were not effective
in reversing the regioselectivity to a useful degree. Potassium
carbonate could also be used as the base, but a minimum of
two equivalents were required for good conversion.
There have been reports on the use of copper catalysis in
biaryl ether formation from halopyridines,12 but none that
address the issue of regioselectivity in 2,4-chloropyridines. The
beneficial effect of pyridine, either as a ligand or as solvent,
has been well-demonstrated in nonpyridyl Ullmann couplings,13
but this example appears to be unique, since the pyridine
appears to play a role even in the absence of copper. It is
possible that pyridine is adding to form a transient pyridinium
species, but literature precedent14 suggests that such activa-
tion would occur at C-4, and would be more likely to enhance
the electrophilicity of that carbon rather than block it. One
reaction was run in which collidine was substituted for
pyridine, and in that instance, almost no conversion was
observed after extensive reaction times. However, a trace of
a product was observed by GC that had a mass consistent
with displacement of one of the chlorides by collidine. If
the pyridinium adduct is forming, the role of the copper is
still not explained. Wright and co-workers have described
the isolation of macrocyclic 2-pyridyl copper complexes,15
which may help explain the directing effect noted in this
case, but would not explain the solvent effect. Finally, an
electron-transfer mechanism cannot be ruled out; the reac-
tions with copper(I) are highly colored and, as noted
previously, occur at a much faster rate than those without it.
The final step of the process was modified to avoid the use
of sodium hydride. Replacement with potassium tert-butoxide
gave clean displacement of the remaining chloride by 3-pen-
tanol, and the solvent was changed from DMSO to DMAc to
reduce potential handling issues. The order of addition and
temperature were found to have a large effect on the conversion.
The base had to be added to a mixture of the starting materials
at a minimum of 120 °C. Running the reaction at lower
temperatures or premixing all the reagents and heating gave
<60% conversion. A THF solution of the base was used to
avoid having to open the reactor under these conditions, with
continuous distillation of the THF from the reaction mixture to
maintain the temperature. The need for these conditions to
obtain good conversion is presumed to be due to steric
hindrance, since any residual methanol not removed after the
isolation of 4 gave the corresponding methoxy-substituted
product at much lower temperatures.
The reason for the changes in selectivity observed is not
entirely clear. There are a few examples reported in the literature
in which alkoxide anions add with high or exclusive selectivity
to the C-4 position of polychloropyridines. The best results are
obtained when the alkoxide is small and the reaction is carried
out at lower temperatures.5 Increasing the size of the nucleophile
tends to lead to mixtures of regioisomers,6 and this is usually
observed with phenols.7 More examples have been reported for
dichloroquinolines8 and other related heterocycles, 9 and in some
cases the regioselectivity can be controlled by reaction condi-
tions,10 but the kinetic selectivity is opposite to that observed
for the pyridines (C-2 > C-411), making it difficult to draw
analogies.
The optimized conditions are shown in Scheme 3. During
scale-up, a silica pad filtration was carried out after the first
coupling to ensure removal of the copper salts. The recovery
for both steps was less than the inherent yield because of the
very high solubility of both 4 and 1 in all solvents screened for
isolation.
In conclusion, a method for regioselective addition of mesitol
to the C-2 position of 2,4-dichloro-3,6-dimethylpyridine has
been described. The use of a copper(I) salt and pyridine as
solvent are key to achieving the desired selectivity.
(5) Tiwari, A.; Riordan, J. M.; Waud, W. R.; Struck, R. F. J. Med. Chem.
2002, 45, 1079.
(6) Brady, J. H.; Wakefield, B. J. Synthesis 1984, 33.
(7) Mack, A. G; Suschitzky, H.; Wakefield, B. J. J. Chem. Soc., Perkin
Trans. 1 1980, 1682.
(8) Osborne, A. G.; Miller, L. A. D. J. Chem. Soc., Perkin Trans. 1 1993,
181.
(12) (a) Hill, A. J; McGraw, W. J. J. Org. Chem. 1949, 14, 783. (b)
Fujiwara, H.; Kitagawa, K. Heterocycles 2000, 53, 409.
(9) Chen, J.; Steglich, W. J. Heterocycl. Chem. 1993, 30, 909.
(10) Jarvest, R. L.; Armstrong, S. A.; Berge, J. M.; Brown, P.; Elder, J. S.;
Brown, M. J.; Copley, R. C. B.; Forrest, A. K.; Hamprecht, D. W.;
O’Hanlon, P. J.; Mitchell, D. J.; Rittenhouse, S.; Witty, D. R. Bioorg.
Med. Chem. Lett. 2004, 14, 3937.
(13) (a) Pello´n, R. F.; Carrasco, R.; Millia´n, V.; Rode´s, L. Synth. Commun.
1995, 25, 1077. (b) Fagan, P. J.; Hauptmann, E.; Shapiro, R.;
Casalnuovo, A. J. Am. Chem. Soc. 2000, 122, 5043.
(14) Schmidt, A.; Mordhorst, T.; Nieger, M. Synthesis 2006, 23, 3987.
(15) Garcia, F.; Hopkins, A. D.; Kowenicki, R. A.; McPartlin, M.; Rogers,
M. C.; Wright, D. S. Organometallics 2004, 23, 3884.
(11) Rowlett, R. J., Jr.; Lutz, R. E J. Am. Chem. Soc. 1946, 68, 1288.
412
•
Vol. 12, No. 3, 2008 / Organic Process Research & Development

Scheme 3. Optimized synthesis
potassium tert-butoxide, 1.57 kg (8.2 mol) of copper(I) iodide,
and 5.81 kg (33.0 mol) of 3. The mixture was heated to reflux
for 2 h and then cooled to 0 °C. The reaction was poured into
a mixture of hexanes (58.3 L) and saturated NH4Cl (116.2 L),
and stirred at room temperature overnight. The layers were
separated, and the organic layer was washed with 2 × 15 L of
1 M NH4OH, 3 × 29 L of 3 N NaOH, 1 × 29 L of 1 N HCl
and 1 × 29 L of water. The organic layer was dried over sodium
sulfate and filtered, washing with hexanes. The filtrate was
concentrated under vacuum to a brown oil and then pushed
through a small plug of silica with hexanes. After removal of
the solvent under vacuum, the residue was stirred in methanol
(8.0 L) to precipitate the product. The slurry was filtered, and
the solids were dried under vacuum to obtain 3.9 kg (43%) of
4. In order to obtain a second crop, the filtrate was concentrated
under vacuum to an oil, and the residue was slurried in minimal
methanol at 0 °C and filtered under vacuum. The solids were
dried to give an additional 1.6 kg (18%) of 4 for an overall
yield of 61%. HPLC analysis showed the product to be 99.1%
pure with 0.6% of the bis-adduct. 1H NMR (CDCl3): δ 6.88 (s,
2H), 6.78 (s, 1H), 2.40 (s, 3H), 2.30 (s, 3H), 2.20 (s, 3H), 2.04
(s, 6H). 13C NMR (400 MHz, CDCl3): δ 161.1, 154.0, 148.9,
145.0, 134.2, 130.7, 129.2, 117.8, 115.4, 23.9, 21.1, 16.8, 12.3.
Anal. Calcd. for C16H18ClNO: C, 69.68; H, 6.58; N, 5.08; Cl,
12.86.. Found: C, 70.03; H, 6.54; N, 5.00; Cl, 12.49.
4-(1-Ethylpropoxy)-3,6-dimethyl-2-(2,4,6-trimethylphe-
noxy)pyridine (1). A vessel was charged with 3.88 kg (14.1
mol) of 4, 2.1 kg (23.8 mol) of 3-pentanol, and 46.4 L of
DMAc. The mixture was heated to 120 °C, and a 1 M solution
of potassium tert-butoxide in THF (34.7 L, 34.7 mol) was added
over 4 h with simultaneous removal of the THF by distillation.
After the addition was complete, the mixture was cooled and
partitioned between hexanes (77.6 L) and water (77.6 L). The
organics were washed with water (2 × 39 L) and dried over
sodium sulfate. The hexanes were displaced with isopropanol
under vacuum to a final volume of 12 L. The resulting slurry
was cooled to -10 °C, and the solids were isolated by filtration.
After drying, 3.9 kg (84%) of 1 was obtained that was 99.4%
pure by HPLC. 1H NMR (400 MHz, CDCl3): δ 6.88 (s, 2H),
6.29 (s, 1H), 4.19 (p, J ) 5.8 Hz, 1H), 2.30 (s, 3H), 2.20 (s,
6H), 2.08 (s, 6H), 1.76-1.69 (m, 4H), 0.97 (td, J ) 7.5, 0.8
Hz, 6H). 13C NMR (400 MHz, CDCl3): δ 164.7, 161.9, 154.0,
149.4, 133.6, 131.1, 129.0, 104.3, 102.5, 80.2, 26.3, 24.8, 21.1,
16.9, 9.7, 8.2. Anal. Calcd. for C21H29NO2: C, 77.02; H, 8.93;
N, 4.28. Found: C, 76.66; H, 8.93; N, 4.22.
Experimental Section
General. All materials were purchased from commercial
suppliers and used without further purification. All reactions
were conducted under an atmosphere of nitrogen unless noted
otherwise. All reactors were glass-lined steel vessels. Reactions
were monitored for completion by removing a small sample
from the reaction mixture and analyzing the sample by HPLC
or GC/MS. HPLC analyses were performed using a Zorbax
SB-CN 4.6 mm × 150 mm column and a gradient mobile phase
consisting of 0.02% phosphoric acid (A) and acetonitrile (B)
(A/B 95/5 to 25/75 at 10 min, hold to 14 min, to 95/5 at 15.5
min, flow 0.5 mL/min, temp 30 °C). GC analyses were
performed on a DBWaxEtr (15 m × 0.25 mm ID × 0.25 µm
film). Proton and carbon NMR spectroscopy was performed
on a Bruker-Spectrospin Avance 400 MHz instrument. Elemen-
tal analyses were performed by Quantitative Technologies Inc.
of Whitehouse, NJ.
2,4-Dichloro-3,6-dimethylpyridine (3). To a vessel were
charged 8.8 kg (63.2 mol) of 2,4-dihydroxy-3,6-dimethylpyri-
dine and 10.0 L (62.8 mol) of N,N-diethylaniline. Phosphorus
oxychloride (41.2 L) was added, allowing the mixture to
exotherm. The mixture was heated to 90-95 °C and held for
2 h, then checked for reaction completion by GC. The excess
phosphorus oxychloride was removed by distillation under
vacuum, and then the crude residue was poured onto water (176
L) with good stirring. Hexanes (176 L) were added, and the
pH of the mixture was adjusted to 2.5 with 50% NaOH. After
separation of the layers, the organics were washed with 2 ×
176 L of 1 N HCl and 2 × 176 L of 1 M NaOH. The solution
was dried over sodium sulfate, and the solvent was removed
under vacuum. The crude residue was pushed through a small
plug of silica with hexanes and isolated by concentration to
obtain 8.0 kg (72%) of the product as a clear amber oil. HPLC
anaylsis showed the material to be 97.6% pure. 1H NMR (400
MHz, CDCl3): δ 7.02 (s, 1H), 2.40 (s, 3H), 2.34 (s, 3H). 13C
NMR (400 MHz, CDCl3): δ 156.4, 151.4, 145.6, 127.6, 123.0,
23.5, 16.4. Anal. Calcd. for C7H7Cl2N: C, 47.76; H, 4.01; Cl,
40.28; N, 7.96. Found: C, 47.81; H, 3.96; Cl, 39.94; N, 7.75.
4-Chloro-3,6-dimethyl-2-(2,4,6-trimethylphenoxy)pyri-
dine (4). A vessel was charged with 4.94 kg (36.3 mol) of 2,4,6-
trimethylphenol, 29 L of pyridine, 4.07 kg (36.3 mol) of
Acknowledgment
We gratefully acknowledge Drs. Yuhpyng Chen, Joel
Hawkins, and Shu Yu and Professors Steve Ley and Dave
Collum for their suggestions and discussions regarding the
chemistry described.
Received for review January 9, 2008.
OP800004Q
Vol. 12, No. 3, 2008 / Organic Process Research & Development
•
413