A practical synthesis of a COX-2-specific inhibitor

Article abstract of DOI:10.1021/jo000870z

A number of synthetic strategies to the Cox-2 specific inhibitor 1 have been described. These studies have led to the identification of a novel pyridine construction using annulation of ketone 2 using a vinamidinium species 29 and ammonia in 97% assay yield. Three approaches to the synthesis of ketone 2 are described that allow for its preparation in large quantities in >65% overall yield from methyl 6-methylnicotinate.

Full text of DOI:10.1021/jo000870z

J . Org. Chem. 2000, 65, 8415-8420
8415
A P r a ctica l Syn th esis of a COX-2-Sp ecific In h ibitor
Ian W. Davies,* J ean-Francois Marcoux, Edward G. Corley, Michel J ournet, Dong-Wei Cai,
Michael Palucki, J immy Wu, Robert D. Larsen, Kai Rossen, Philip J . Pye, Lisa DiMichele,
Peter Dormer, and Paul J . Reider
Department of Process Research, Merck & Co., Inc., P.O. Box 2000, Rahway, New J ersey 07065
ian_davies1@merck.com
Received J une 7, 2000
A number of synthetic strategies to the Cox-2 specific inhibitor 1 have been described. These studies
have led to the identification of a novel pyridine construction using annulation of ketone 2 using
a vinamidinium species 29 and ammonia in 97% assay yield. Three approaches to the synthesis of
ketone 2 are described that allow for its preparation in large quantities in >65% overall yield from
methyl 6-methylnicotinate.
In tr od u ction
Nonsteroidal antiinflammatory drugs (NSAIDs) used
for the treatment of inflammatory conditions act by
inhibition of cyclooxygenase (COX), the first enzyme
involved in the biosynthesis of prostaglandins, prosta-
cyclins, and thromboxanes from arachidonic acid. The
major COX isozyme, COX-1, is expressed as a constitutive
enzyme and is involved in homeostasis of the gastrointes-
tinal (GI) tract (in addition to other functions).1 The
discovery2 of an inducible COX isozyme, commonly
referred to as COX-2 and expressed principally in inflam-
matory tissue, has led several groups to search for
selective inhibitors of COX-2.3 The rationale behind these
investigations is that a specific COX-2 inhibitor will
greatly improve the side-effect profile, including gastric
ulceration, that is commonly associated with the chronic
use of traditional NSAIDs. Recently, a series of novel
2-pyridinyl-3-(4-methylsulfonyl)phenylpyridines has been
evaluated by Merck for their ability to inhibit the
isozymes of cyclooxygenase, COX-1, and COX-2. Optimal
COX-2 inhibition was observed by introduction of a
substituent at C5 of the central pyridine. 5-Chloro-3-(4-
methylsulfonyl)phenyl-2-(2-methyl-5-pyridinyl)pyridine 1
was identified as a very potent and specific COX-2
inhibitor that may provide therapeutically useful alter-
natives to traditional NSAIDs with a greater GI safety
profile.3
proach was initiated. We envisioned that these com-
pounds may be assembled by construction of the central
pyridine ring (Scheme 1) with the introduction of the C-5
substituent in a single step. This disconnection would
lead to ketone 2, a three-carbon annulating agent 3, and
ammonia.
Previously, the efficient construction of trisubstituted
pyridines via the annulation of ketones with vinami-
dinium salts was described in a Letter.4 In this paper, a
number of synthetic approaches to the ketosulfone 2 and
full details of the development of the efficient practical
annulation strategy for the construction of the central
pyridine ring are disclosed.
Resu lts a n d Discu ssion
Ap p r oa ch es to Ketosu lfon e 2. Three of the ap-
proaches that have been examined are outlined below
and have relied on the construction of the central C-C
single bond of the ketosulfone. Our strategy has required
the use of a preformed 6-methyl-substituted pyridine
ring, and with the ready availability from the Lonza
nicotinamide process, methyl 6-methyl nicotinate 4 was
the most obvious choice of starting material.5 Two ap-
proaches directly involve the 6-methylnicotinate as an
electrophilic component whereas the third uses the
6-methylpyridine as nucleophilic component after ap-
propriate derivatization.
Previous synthetic strategies to these diaryl pyridines
have employed transition metal cross-couplingsSuzuki
or Stillesof preformed pyridines and were very effective
at preparing a wide range of substrates for preliminary
biological evaluation.3 With the identification of 1 as a
development candidate, a more focused synthetic ap-
* Corresponding author. Tel: 732 594 3174. Fax: 732 594 1499.
(1) For a review see: Griswold, D. E.; Adams, J . L. Med. Res. Rev.
1996, 16, 181.
(1) Hor n er -Wittig Ap p r oa ch . This approach relies
on the modified Horner-Wittig reaction developed by
Zimmer.6 Coupling of the N,P-acetal 6 and the com-
mercially available 4-methanesulfonylbenzaldehyde gen-
(2) (a) Xie, W.; Chipman, J . G.; Robertson, D. L.; Erickson, R. L.;
Simmons, D. L. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 2692. (b)
Kujubu, D. A.; Fletcher, B. S.; Varnum, B. C.; Lim, R. W.; Herschman,
H. R. J . Biol. Chem. 1991, 226, 2692. (c) Masferrer, J . L.; Seibert, K.;
Zweifel, B.; Neddleham, P. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 3917.
(d) Hla, T.; Neilson, K. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 7384.
(3) Friesen, R. W.; Brideau, C.; Chan, C. C.; Charleson, S.; De-
scheˆnes, D.; Dube´, D.; Ethier, D.; Fortin, R.; Gauthier, J . Y.; Girard,
Y.; Gordon, R.; Greig, G. M.; Riendeau, D.; Savoie, C.; Wang, Z.; Wong,
E.; Visco, D.; Xu, J . L.; Young, R. N. Bioorg. Med. Lett. 1998, 8, 2777.
(4) Marcoux, J .-F.; Corley, E. G.; Rossen, K.; Pye, P.; Wu, J .; Robbins,
M. A.; Davies, I. W.; Larsen, R. D.; Reider, P. J . Org. Lett. 2000, 2,
2339.
(5) Grayson, J . I. EP 0128279A2, 1984.
10.1021/jo000870z CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/19/2000

8416 J . Org. Chem., Vol. 65, No. 25, 2000
Davies et al.
Sch em e 1
Sch em e 3a
Sch em e 2a
a
Reagents: (a) HNMe(OMe), i-PrMgCl, toluene; (b) DIBAL-H,
toluene.
to the pyridine, ester, or Weinreb amide were minimized
to <0.1% under these conditions. The replacement of
tetrahydrofuran by toluene aided extractions during
isolation and allowed for the introduction of a through
process. The reaction was quenched into 1 equiv of 10%
aqueous acetic acid which proved to be an excellent
system for extraction of the magnesium salts since Mg-
(OAc)2 is a very soluble salt (390 g/L) in contrast to the
complex (NH4)MgCl3 which has low solubility.9 The pH
of the quench is also weakly acidic and avoids the
precipitation of magnesium hydroxide. With these modi-
fications the Weinreb amide was obtained as a 25 wt %
solution in toluene in >95% yield (Scheme 3).
a
Reagents: (a) aniline, diphenyl phosphite, IPAC; (b) potassium
Reduction of the intermediate Weinreb amide 8 was
achieved in toluene using 20 wt % DIBAL-H at <-15 °C
to give the aldehyde 5 in 92% yield. The reaction was
accompanied by <5% over-reduction. Isolation of the
aldehyde was achieved by quenching excess reagent with
1 equiv of ethyl acetate and workup using aqueous (L)-
tartaric acid. This procedure led to isolation of the desired
aldehyde with varying amounts of the aminal 9, which
could be used in the next step without complication. The
solution of aldehyde 5 was solvent switched to a 30 wt %
solution in IPAC for use in the next step.
Since the N,P-acetal 6 is the first crystalline interme-
diate in the synthesis, it was important that impurities
from the previous steps were adequately rejected. After
much experimentation, the optimal solvent for isolation
of the N,P-acetal was isopropyl acetate/n-heptane. This
led to the selection of IPAC as the solvent for the acetal
formation. Treatment of a solution of the aldehyde in
IPAC with 1.2 equiv of aniline at 60 °C led to the
formation of the intermediate imine. The aminal 9 was
also converted to the imine under these conditions.
Addition of 1.7 equiv of technical grade diphenyl phos-
phite (∼85% phosphite, ∼15% phenol) over 30 min at 60-
65 °C formed the N,P acetal in >92% assay yield. The
reaction mixture was seeded with 6 and the product
crystallized by the addition of 1.1 volumes of n-heptane
over 2 h followed by cooling to 15 °C. The crystalline N,P-
acetal was isolated by filtration in >99A% by HPLC in
87% isolated yield. The control of the aniline and phenol
levels in the N,P-acetal were critical to ensure the yield
and quality of ketosulfone obtained in the next step.
A wide range of bases, solvents and temperatures were
examined for the conversion of N,P-acetal 6 to ketosul-
fone 2. The original conditions reported for the reaction
tert-butoxide, IPA/THF; (c) 2 N HCl.
erates the enamine 7 which after hydrolysis leads to the
desired ketosulfone 2 (Scheme 2). The N,P-acetal is
derived from the nicotinaldehyde 5, and in order for this
approach to be successful a straightforward synthesis of
this aldehyde was required.
For initial probe experiments the 6-methyl methylni-
cotinate 4 was reduced to the alcohol using LAH and
oxidized to the aldehyde 5 using MnO2 in 91% overall
yield.7 This pathway was not suitable for large scale
production of 5. Reduction of the ester 4 using 1 equiv of
DIBAL-H in methylene chloride, toluene, or ethereal
solvents led to clean formation of the alcohol in 50%
conversion independent of the order of addition or quench
method used. With the propensity for over-reduction of
the ester, the reactivity of the Weinreb amide 8 was
investigated. The direct conversion of esters to their
Weinreb amides has previously been described by Wil-
liams.8 In the published procedure, a slurry of N,O-
dimethylhydroxylamine hydrochloride and appropriate
ester in THF was treated with isopropylmagnesium
chloride followed by aqueous ammonium chloride quench
to give the Weinreb amide in high yield. Slight modifica-
tions were introduced to make this procedure more
efficient. By reaction of 1.6 equiv of N,O-dimethylhy-
droxylamine as the free-base in toluene at <-7 °C, 1.4
equiv of isopropylmagnesium chloride could be used in
place of the 2.8 equiv required for the HCl salt. The
presence of byproducts resulting from i-PrMgCl addition
(6) (a) Zimmer, H.; Bercz, J . P. J ustus Liebigs Ann. Chem. 1965,
686, 107-114. (b) Bunting, J . W.; Stefanidis, D. J . Am. Chem. Soc.
1988, 110, 4008-4017.
(7) Nishikawa, Y.; Shindo, T.; Ishii, K.; Nakamura, H.; Kon, T.; Uno,
H. J . Med. Chem. 1989, 32, 583.
(8) Williams, M. J .; J obson, R. B.; Yasuda, N.; Marchesini, G.;
Dolling, U. H.; Grabowski, E. J . J . Tetrahedron Lett. 1995, 36, 5461.
(9) Seidell, A Solubilities of Inorganic and Metal Organic Compound;
Van Nostrand: New York, 1940.

Synthesis of a COX-2-Specific Inhibitor
J . Org. Chem., Vol. 65, No. 25, 2000 8417
by Zimmer5 utilized methanol/KOH at -40 °C which we
have already reported can lead to a significant side
reaction via saponification of the phosphonate.10 The use
of CsCO3 in THF/IPA followed by HCl hydrolyis of the
enamine led to a smooth reaction to give ketosulfone in
85% isolated yield. However this procedure requires the
reaction of an excess of the insoluble CsCO3 with the N,P-
acetal which in turn is coupled with a sparingly soluble
aldehyde. The reaction rate was shown to be a function
of the particle size of the cesium carbonate. This proce-
dure was troublesome upon scaleup. An additional prob-
lem was the generation of the alcohol impurities 10 and
11 at up to 1 A%. The bis-sulfone impurity 12 was also
present in up to 0.5 A%. These ketone impurities carried
solvents were screened. In our hands, the use of THF led
to almost exclusive formation of the ethane dimer 16,12
while the use of other ethereal solvents13 failed to give
any Grignard formation. However, similar results to
those using diethyl ether were obtained by preparing the
Grignard reagent in toluene using 1.5 equiv of magne-
sium and addition of the chloride dissolved in 2 equiv of
tetrahydrofuran at <35 °C. Addition of the Grignard
reagent (after decanting off the excess magnesium) to the
Weinreb amide at <-15 °C gave the ketosulfide 15 in
80% isolated yield at 1 mol scale.
Oxidation of the sulfide 15 to ketosulfone 2 was
achieved very effectively using Oxone in ethyl acetate.14
However, due to the polarity of the ketosulfone, it proved
difficult to obtain a pure sample free of inorganics.
Oxidation using acetic acid/hydrogen peroxide at 100 °C
was very sluggish. Tungstate-catalyzed oxidation proved
to be effective in ODCB or methanol using aqueous
hydrogen peroxide as the stoichiometric oxidant. Oxida-
tion of the ketosulfide in the presence of sulfuric acid at
60 °C minimized the formation of N-oxide 17 to <0.3 A%
via in situ protection of the pyridine. Ketosulfoxide 18
was readily isolable at incomplete conversion. Methane-
sulfonic acid can also be used in place of sulfuric acid.
An excess of hydrogen peroxide was required to drive the
reaction to completion due to a competitive oxidation of
methanol under these conditions. The ketosulfone was
isolated in 89% yield by filtration after cooling to ambient
and washing with water.
forward to pyridines in the annulation reaction with
essentially quantitative conversion and were difficult to
reject thus compromising the purity of 1. For this reason
a method to minimize their formation in the ketosulfone
step was investigated. Optimal conditions involved ad-
dition of 1.05 equiv of potassium tert-butoxide over 1 h
to a slurry of N,P-acetal in THF/IPA (0.7:1) at 25 °C. The
enamine intermediate was hydrolyzed by addition of 4
equiv of 2 N HCl to give the ketosulfone in 94% assay
yield. The formation of the alcohol impurities was es-
sentially inhibited under these conditions. The ketosul-
fone product was isolated by adjusting the pH of the
solution to ∼5.5 with 5 N NaOH at 45 °C. The ketosulfone
was isolated by filtration following cooling to -10 °C and
washing with IPA/water (1:1) to give an 87% isolated
yield of 2 with >99 A% purity by HPLC.
The Horner-Wittig strategy led to generation of ke-
tosulfone in four steps and 65% overall yield from methyl
6-methylnicotinate 4.
(2) Gr ign a r d Ap p r oa ch . Since the Weinreb amide 8
is an intermediate in the synthesis of nicotinaldehye 5,
addition of a nucleophilic organometallic reagent to 8
might provide direct access to the ketosulfone.
The Grignard/oxidation route afforded ketosulfone 2
in three step and 68% overall yield from methyl 6-me-
thyllnicotinate 4.
(3) Cla isen Ap p r oa ch . An alternative construction
of ketosulfone 2, again relying on the nicotinate 4 as the
source of the pyridine, utilizes a Claisen condensation of
4 with 20 or 21. The substituted phenyl acetic acid 20 or
ethyl ester 21 as the required nucleophilic partner have
been previously described.14 However, due to the ready
availability of the ketone 19, the acid 20 was conveniently
prepared using the Willgerodt-Kindler reaction.16 The
(12) This is in contrast to a literature report where coupling with
N-methyl-3,4-lutidinium iodide was reported to occur in THF with a
crude yield of 44%: Hori, M.; Ban, M.; Imai, E.; Iwata, N.; Suzuki, Y.;
Baba, Y.; Morita, T.; Fujimura, H.; Nozaki, M.; Niwa, M. J . Med. Chem.
1985, 28, 1660.
(13) J ones, R. V. H.; Ewins, R. C.; Mellor, R. M. US 5,242,625, 1993.
(14) For tungstate oxidation of a related COX-2 inhibitor, see: Tan,
L.; Chen, C.; Larsen, R. D.; Verhoeven, T. R.; Reider, P. J . Tetrahedron
Lett. 1998, 39, 3961. For Oxone oxidation of a related Cox-2 inhibitor
see: Xu, F.; Zhao, D.; Tillyer, R. D.; Grabowski, E. J . J .; Reider, P. J .
Tetrahedron 1999, 55, 6001.
The use of 4-thiomethylbenzylmagnesium chloride 14,
derived from the chloride in diethyl ether, has been
employed by this group as an intermediate in the
synthesis of Clinoril.11 Reaction of a 2 M solution of the
Grignard 14 in ether with the Weinreb amide in toluene
led to the ketone in 82% isolated yield. To avoid the use
of diethyl ether in the Grignard formation, a variety of
(10) J ournet, M.; Cai, D.-W.; Larsen, R. D.; Reider, P. J . Tetrahedron
Lett. 1998, 39,1717.
(11) Pines, S. H.; Czaja, R. F.; Abrahamson, N. L.. J . Org. Chem.
1975, 40, 1920.
(15) Forrest, H. S.; Fuller, A. T.; Walker J . J . Chem. Soc. 1948, 1501.
Forrest, H. S.; Fuller, A. T.; Walker J . J . Chem. Soc. 1948, 1504.
(16) Willgerodt, C. Chem. Ber. 1887, 20, 2467. Willgerodt, C. Chem.
Ber. 1888, 21, 534. Kindler, K. Liebig’s Annalen 1923, 431, 193.

8418 J . Org. Chem., Vol. 65, No. 25, 2000
Davies et al.
sa tion .17 When a mixture of chloromalonaldehyde (3
equiv),18 22, ammonium acetate (9 equiv), and ketosul-
fone 2 were heated with propionic acid at 125 °C, pyridine
1 was produced in 62% assay yield together with a the
furan impurity 23 in 15% assay yield. Isolation by
chromatography led to the desired pyridine 1 in 49% yield
and the furan 23 in 11% yield. In addition to the furan,
the 3,5-dichloropyridine 24 was observed in up to 5%
assay yield. The generation of these two impurities
suggests that decarbonylation of a chloromalonaldehyde
derivative is a significant reaction pathway under these
conditions. The reaction of aniline derivative 25, an
isolable intermediate in the Dieckmann synthesis of
chloromalonaldehyde, behaved similarly in the reaction
to give the pyridine 1 in 47% assay yield.
acid was obtained in 57% yield, and the ester 21 was
prepared using Fischer esterification in 99% yield.
Reaction of the ester 21 with methyl nicotinate 4 using
potassium tert-butoxide in tert-butyl alcohol as base
followed by acid hydrolysis/decarboxylation only led to
small amounts of the desired ketsoulfone 2. This was
presumably due to the ease of retro-reaction in this
system. The use of Ivanoff conditions, i.e., the magnesium
dianion in THF (formed using 2 equiv of tert-butylmag-
nesium chloride) at 50 °C proved to be more success-
ful; however, in this case the conversion was limited to
< 50%. This problem was overcome by the addition of
an extra 1 equiv of tert-butyl magnesium chloride during
the course of the reaction. Therefore, by using a total of
3 equiv of tert-butylmagnesium chloride the ketosulfone
2 could be obtained in 58% yield directly from nicotinate
4.
Careful monitoring by HPLC indicated that aminoac-
rolein 26 was an intermediate in the reaction. The
aminoacrolein was conveniently prepared from the vi-
nylogous isopropyl ester 27 in 80% overall yield starting
with chloromalonaldehyde. When the annulation reaction
was performed with the isolated aminoacrolein in the
absence of ammonium acetate, the level of furan 23 was
significantly reduced to <5%. Four equivalents of ami-
noacrolein was still required to drive the reaction to
completion. By addition of a strong acid the formation of
the furan could be further minimized. Optimal conditions
utilized 2 equiv of methanesulfonic acid in a mixture of
toluene/propionic acid where the levels of furan could be
reduced to <1% with 72% assay yield. However, the
reaction was still plagued by large amounts of polymeric
material and suffered from the serious disadvantage that
an excess of aminoacrolein was required. Further study
revealed that the reaction of aminoacrolein to form the
head-to-head dimer 28 was a competing side reaction
under these conditions. While kilogram quantities of
pyridine 1 could be prepared with this approach, it was
P yr id in e An n u la tion of Ketosu lfon e. The key dis-
connection in the synthesis of 1 that we wished to exploit
was the annulation of ketosulfone 2 with an appropriate
three-carbon synthon 3 and ammonia (Scheme 1). Two
broad classes of reaction have been investigated using
acid- or base-induced enolization of the ketosulfone and
subsequent reaction with a derivative of chloromalonal-
dehyde 22.
(17) Breitmaier, E.; Bayer, E., Angew. Chem., Int. Ed. Engl. 1969,
8, 765.
(18) Dieckmann, W.; Platz, L. Ber. Deut. Chem. Ges. 1904, 37, 4638.
Chloromalonaldehyde is commercially available from PPG Industries.
Acid -P r om oted An n u la tion /F r ied la n d er Con d en -

Synthesis of a COX-2-Specific Inhibitor
J . Org. Chem., Vol. 65, No. 25, 2000 8419
Sch em e 4
apparent that a more efficient process was required for
the production of even larger amounts of 1.19
(2) Ba se-P r om oted An n u la tion . Alternatively reac-
tion of the enolate of ketone 2 with a variety of three
carbon electrophiles was examined. 2,3-Dichloroacrolein
3020 was prepared by treatment of chloromalonaldehyde
with oxalyl chloride in toluene and catalytic DMF fol-
lowed by distillation in 80% yield. Reaction of the lithium
enolate of ketosulfone with dichloroacrolein 30 in THF
followed by reaction with ammonim acetate or anhydrous
ammonia at reflux led to the formation of pyridine 1 in
up to 58% assay yield. The reaction could not be opti-
mized further, and this approach was discontinued due
to expected difficulty in handling dichloroacrolein at large
scale.
tert-butoxide or LDA decreased the yields significantly
and resulted in the recovery of large quantities of
unreacted ketosulfone. The desired pyridine 1 was iso-
lated in 94% yield following chromatography. The pKa
of 1 was determined to be 4.524 making it more conve-
nient on a large scale to isolate the pyridine as the
p-toluenesulfonic acid salt in 91% yield without resorting
to chromatography.
Su m m a r y
We have described syntheses which allow for the
preparation of 1 in four or five steps with >60% overall
yield from methyl 6-methylnicotinate. These studies have
resulted in the discovery of a novel pyridine construction
involving the annulation of ketosulfone and CDT-
phosphate. This reaction has significant advantages over
more traditional acid-promoted reactions and should
provide broad applicability for the synthesis of substi-
tuted pyridines.
Vinamidinium salts have been used in the synthesis
of nitrogen-containing heterocycles, e.g., pyrroles and
pyrimidines. Carbonyl enolates are known to add to
vinamidinium salts to produce dienaminones,21 and eno-
lates of substituted aryl methyl ketones are alkylated by
vinamidinium salts.22 In the latter case the formation of
pyridine rings was achieved in a three-step process in
36-46% yield via the formation of the corresponding
pyrylium salts and reaction with ammonium acetate. We
have described the preparation of several vinamidinium
hexafluorophosphate salts and recently reported that
these salts are suitable three-carbon synthons that react
with ketone enolates to form 2,5-disubstituted-3-arylpy-
ridines in good to excellent yield.4
Exp er im en ta l Section
Melting points were determined are uncorrected. Elemental
analyses were performed by Quantitative Technologies, Inc.,
Whitehouse, NJ . Water content was determined by Karl
Fischer titration.
To this end 2-chloro-N,N-dimethylaminotrimethinium
hexafluorophosphate salt 2923 (1.05 equiv) was reacted
with ketone 2 (1.00 equiv) in the presence of an equimolar
amount of t-BuOK in THF at ambient temperature
(Scheme 4). Inverse quench of the resulting adduct 31
into a mixture of 7 equiv of HOAc and 0.75 equiv of TFA
in THF at <30 °C led to the putative intermediate 32.
Ring closure of the pyridine ring occurred upon heating
at reflux in the presence of an excess of aqueous am-
monium hydroxide to give the pyridine 1 in 97% assay
yield. Replacement of the potassium base with sodium
1-(6-Me t h yl-3-p yr id in yl)-1-(a m in op h e n yl)m e t h a n o-
d ip h en ylp h osp h on ic Acid 6. To a solution of methyl 6-me-
thylnicotinate (21.56 g, 0.143 mol) and N,O-dimethylhydrox-
ylamine (13.90 g, 0.30 mol) in toluene (150 mL) at -10 °C was
added isopropylmagnesium chloride (110 mL, 2 M in THF)
over 2.5 h. The reaction mixture was aged for 30 min and then
transferred to 10% aqueous acetic acid (126 mL) at 5 °C. The
layers were separated, and the aqueous layer was extracted
with toluene (×2). The layers were concentrated with the
concomitant addition of toluene to reduce the KF to <200 µg/
mL to give a ∼30 wt % solution of 8 in toluene. The solution
was cooled to - 20 °C and DIBAL-H (99 mL, 1.5 M in toluene,
1.1 equiv) was added over 30 min allowing the temperature
to reach -5 °C over the course of the addition. Excess reagent
was quenched by the addition of ethyl acetate (15 mL). The
reaction is aged for 30 min at -10 °C to 0 °C and then added
to a 20% tartaric acid solution maintaining the temperature
<30 °C. The mixture is aged for 1 h at 20-25 °C with efficient
(19) One further acid-promoted approach was briefly investigated.
Reaction of the vinamidinium salt 29 with the ketone 2 using propionic
acid as solvent in the presence of ammonium acetate at 100 °C led to
the formation of pyridine in 30% assay yield. Further optimization of
this reaction was not undertaken due to the success of a related base-
promoted reaction.
(20) J aehnisch, K.; Duczek, W. Synthesis 1992, 10. 935. J aehnisch,
K.; Seeboth, H.; Duczek, W. Chem. Abstr. 88, 492278. Patent Appl.
DD 86-294822. Brackeen, M.; Howard, H. R. EP 0352959.
(21) Nair, V.; Cooper, C. S. J . Org. Chem. 1981, 46, 4759.
(22) Pavlyuchenko, A. I.; Smirnova, N. I.; Kovshev, E. I. Khim.
Geterotsikl. Soedin. 1985, 10, 1392.
(23) Davies, I. W.; Marcoux, J .-F.; Wu, J .; Corley, E. G.; Robbins,
M. A.; Palucki, M.; Tsou, N.; Ball, R. G.; Dormer, P.; Larsen, R. D.;
Reider, P. J . J . Org. Chem. 2000, 65, 4571.
(24) pKa was measured in a range of methanol/water mixtures and
extrapolated to 100% water.

8420 J . Org. Chem., Vol. 65, No. 25, 2000
Davies et al.
stirring. The pH is adjusted to 8.0 with 50% sodium hydroxide,
and the mixture is aged with stirring for 30 min. The layers
are separated, and the aqueous layer is extracted with
isopropyl acetate. The organic layers are concentrated succes-
sively with the concomitant addition of IPAC to give a ∼30 wt
% solution aldehdye 5 with KF < 200 µg/mL. IPAC (160 mL)
and aniline (12.8 g, 0.137 mol) were added, and the mixture
was heated to 60 °C. Diphenyl phosphite (60 g, 86 wt %
technical grade, 0.212 mol) was added over 1 h. The mixture
was seeded with N,P-acetal 6, and heptane (180 mL) was
added over 1 h. The mixture was cooled to 15 °C, and the
product isolated by filtration. The cake was slurry washed with
water followed by IPAC/heptane (1:1) and dried to give N,P-
acetal as a colorless solid (46.5 g) in 76% yield: mp 135-137
°C; 1H NMR (400 MHz CDCl3) δ 8.69 (t, 1H, J ) 2.3 Hz), 7.26
(m, 4H), 7.15 (m, 7H), 6.98 (m, 2H), 6.77 (t, 1H, J ) 7.4 Hz),
6.64 (dd, 2H, J 1)8.6 Hz, J 2)0.9 Hz), 5.17 (d, 1H, J ) 24.7
Hz), 4.89-5.09 (br, 1H), 2.55 (d, 3H, J ) 2.0 Hz); 13C NMR
(100 MHz CDCl3) L 158.4, 158.3, 150.2, 150.1, 150.0, 149.9,
148.7, 148.7, 145.4, 145.2, 136.0, 135.9, 129.7, 129.7, 129.4,
129.3, 128.0, 125.50 125.3, 123.5, 123.5, 120.5, 120.5, 120.2,
120.1, 119.2, 115.5, 114.0, 54.2, 52.7, 23.9. Anal. Calcd for
C25H23N2O3P: C, 69.8; H, 5.4; N, 6.5. Found: C, 69.8; H, 5.2;
N, 6.3.
1-(6-Meth yl-3-p yr id in yl)-2-[4-(m eth ylsu lfon yl)p h en yl]-
eth a n on e 2. P r oced u r e A: To a suspension of N,P-acetal 6
(44.38 g, 0.100 mol) and sulfonylbenzaldehyde (22.14 g, 0.113
mol) in tetrahydrofuran (150 mL) and isopropyl alcohol (300
mL) at 25 °C was added dropwise potassium tert-butoxide (71
mL, 1.6 M in THF, 0.113 mol) over 2 h. The resulting solution
was aged 30 min and was treated with 2 N HCl (200 mL).
The homogeneous reaction mixture was aged 1 h and was
heated to 45 °C. Sodium hydroxide (5 N) was added dropwise
to the reaction mixture (71 mL) to pH 5.0-5.3. The reaction
mixture was aged 1 h at 45 °C and cooled to -15 °C over 3 h.
The product was filtered, and the cake was washed with cold
IPA/water (1:1) and water and dried by a nitrogen sweep to
give ketosulfone 2 (24.91 g) as an off-white solid in 86%
isolated yield.
Hz, J 2)2.3 Hz), 7.2 (m, 5H), 4.21 (s, 2H), 2.60 (s, 3H), 2.44 (s,
3H); 13C NMR (100 MHz CDCl3) δ 196.0, 163.3, 149.6, 137.2,
136.1, 130.5, 129.8, 129.2, 127.0, 123.3, 45.1, 24.7, 15.8. Anal.
Calcd for C15H15NOS: C, 70.01; H, 5.87; N, 5.44. Found: C,
69.96; H, 5.82; N, 5.37.
1-(6-Meth yl-1-oxid o-3-p yr id in yl)-2-[4-(m eth ylsu lfon yl)-
p h en yl]eth a n on e 17: mp 180-183 °C; 1H NMR (400 MHz
CDCl3) δ 8.85 (s, 1H), 7.92 (d, 2H, J ) 8.3 Hz), 7.73 (d, 1H, J
) 8.0 Hz), 7.42 (m, 3H), 4.33 (s, 2H), 3.05 (s, 3H), 2.58 (s, 3H);
13C NMR (100 MHz CDCl3) δ 192.5, 153.7, 139.7, 139.5, 139.0,
132.7, 130.6, 127.8, 126.5, 124.3, 45.1, 44.4, 18.03. Anal. Calcd
for C15H15NO4S: C, 59.00; H, 4.95; N, 4.59. Found: C, 59.11;
H, 4.98; N, 4.52.
1-(6-m eth yl-3-p yr id in yl)-2-[4-(m eth ylsu lfin yl)p h en yl]-
eth a n on e 18: mp 133-134 °C;1H NMR (400 MHz CDCl3) δ
9.12 (s, 1H), 8.16 (dd, 1H, J 1)8.1 Hz, J 2)2.0 Hz), 7.62 (d, 2H,
J ) 8.1 Hz), 7.43 (d, 2H, J ) 8.1), 7.28 (d, 1H, 8.3 Hz), 4.34 (s,
2H), 2.71 (s, 3H), 2.63 (s, 3H); 13C NMR (100 MHz CDCl3) δ
195.3, 163.7, 149.4, 144.53, 137.0, 136.1, 130.5, 123.9, 123.4,
45.1, 43.8, 24.7. HRMS Found: 273.10, Calcd for C15H15
NO2S: 273.08.
-
5-Ch lor o-3-(4-m eth a n esu lfon ylp h en yl)-6′-m eth yl[2,3′]-
bip yr id in yl 1. To a suspension of ketosulfone 2 (7.24 g,
0.025mol) in dry THF (50 mL) at 0 °C was added dropwise a
20 wt % solution of t-BuOK in THF (0.0263 mol). The yellow
slurry was stirred at room temperature for 45 min, and the
hexafluorophosphate salt 29 (8.05 g, 0.0263 mol) was added
in one portion. The resulting mixture was stirred at room
temperature for 45 min and transferred dropwise with a
cannula under nitrogen pressure to a solution of acetic acid
(0.175 mol) and TFA (0.02 mol) in THF (25 mL) at 25-30 °C.
The mixture was stirred for 45 min, and ammonium hydroxide
(15 mL, 0.25) was added in one portion. The resulting dark
solution was heated at reflux for 3 h and cooled to ambient.
The phases are cut, and the organic layer was concentrated
under reduced pressure. Column chromatography on silica gel
eluting with ethyl acetate/hexane gave the bipyridine as a
colorless solid (8.42 g) in 94% yield. Rf 0.5 IPA/hexane/
triethylamine (69:30:1); DSC onset 136.7 °C, peak 138.1 °C;
1H NMR (400 MHz CDCl3) δ 8.69 (d, 1H, J ) 2.3 Hz), 8.36 (3,
1H, J ) 2.2 Hz), 7.88 (d, 2H, J ) 8.4 Hz), 7.72 (d, 1H, J ) 2.3
Hz), 7.54 (dd, 1H, J 1)8.0 Hz, J 2)2.3 Hz), 7.38 (d, 2H, J ) 8.5
Hz), 7.07 (d, 1H, J ) 8.0 Hz), 3.06 (s, 3H), 2.51 (s, 3H); 13C
NMR (100 MHz CDCl3) δ 158.4, 152.2, 149.7, 148.3, 143.7,
140.1, 137.9, 137.2, 135.18. 131.1, 130.0, 130.3, 127.8, 122.7,
44.4, 24.1. Anal. Calcd for C18H15ClN2O2S. C, 60.25; H, 4.21;
N, 7.81. Found: C, 60.30; H, 4.25; N, 7.85.
P r oced u r e B: To a solution of sulfide 15 (270 g, 1.05 mol)
and sulfuric acid (2 N, 20 mL) in methanol (2700 mL) at 55
°C was added a solution of sodium tungstate (6.0 g, 0.02 mol)
in water (200 mL). Hydrogen peroxide (380 mL, 30% aq) was
added over 2 h. Water (3000 mL) was added over 0.5 h, and
the mixture was cooled to ambient. The product was isolated
by filtration, washed with water, and dried to give ketosulfone
2 as a colorless solid (269 g) in 89% isolated yield. DSC peak
1
185.6 °C; H NMR (400 MHz CDCl3) δ 9.10 (s, 1H), 8.15 (dd,
1H, J 1)8.1 Hz, J 2)2.3 Hz), 7.89 (d, 2H, J ) 8.4 Hz), 7.45 (d,
2H, J ) 8.3 Hz), 7.28 (d, 1H, J ) 8.1 Hz), 4.37 (s, 2H), 3.03 (s,
3H), 2.63 (s, 3H); 13C NMR (100 MHz CDCl3) L 194.8, 163.9,
149.4, 140.1, 139.3, 136.0, 130.6, 129.0, 127.6, 123.4, 45.1, 44.4,
24.7. Anal. Calcd for C15H15NO3S: C, 62.26; H, 5.23; N, 4.84.
Found: C, 62.27; H, 5.17; N, 4.74.
1-(6-Meth yl-3-p yr id in yl)-3-(4-m eth a n esu lfon ylp h en yl)-
fu r a n 23: mp 203 °C; 1H NMR (400 MHz DMSO-d6) δ 8.53 (d,
1H, J ) 2.4 Hz), 7.90 (m, 2H0, 7.73 (dd, 1H, J 1)8.3 Hz, J 2)2.4
Hz), 7.69 (d, 1H, J ) 2.0), 7.62 (m, 2H), 7.23 (d, 1H, J ) 8.3
Hz), 6.75 (d, 1H, J ) 2.0), 3.08 (s, 3H), 2.52 (s, 3H); (100 MHz
CDCl3) δ 158.9, 148.0, 147.1, 144.5, 140.8, 140.1, 135.7, 130.1,
128.7, 125.2, 124.3, 122.8, 114.4, 44.5, 24.0. HRMS Calcd for
1-(4-(Met h ylsu lfon yl)p h en yl) 2-[4-(m et h ylsu lfon yl)-
1
p h en yl]eth a n on e 12: mp 151-153 °C; H NMR (300 MHz,
C
17H16NO3S: [M + H]+ ) 314.085. Found: 314.087.
d6-DMSO) L 8.29 (d, J ) 8 Hz, 2H), 8.10 (d, J ) 8 Hz, 2H),
7.89 (d, J ) 8 Hz, 2H), 7.55 (d, J ) 8 Hz, 2H), 4.67 (s, 2H),
3.30 (s, 3H), 3.22 (s, 3H). Anal. Calcd for C16H16O5S2: C, 54.53;
H, 4.58. Found: C, 54.95; H, 4.58.
2-Ch lor oa m in oa cr olein 26. A solution of chloromalonal-
dehyde (110.5 g, 1.0 mol) in 2-propanol (150 mL) was concen-
trated to a dark brown oil which was added to a solution of
ammonium hydroxide (46 mL) in 2-propanol (200 mL which
resulted in an exotherm to 40 °C. The mixture was stirred for
12 h at ambient temperature, and the product was isolated
by filtration. The cake was washed with 2-propanol and dried
to give the aminoacrolein 26 (84 g) as an off white solid in
80% isolated yield: mp 150-152 °C; 1H NMR (300 MHz CDCl3)
L 9.13 (s, 1H), 7.55 (s, 1H); Anal. Calcd C3H4ClNO: C, 34.15;
H, 3.82; N, 13.27. Found: C, 34.22; H, 3.73; N, 13.11.
2-Ch lor oa m in oa cr olein d im er 28: mp 108-109 °C; 1H
NMR (200 MHz CD3CN) δ 9.27 (s, 2H), 7.92 (s, 2H), 2.14 (s,
1H); 13C NMR (100 MHz DMSO-d6) δ 222.0, 184.4, 156.4. Anal.
Calcd for C6H5Cl2NO2: C, 37.14; H, 2.60; N, 7.22. Found: C,
37.15; H, 2.50; N, 7.07.
1-(6-Meth yl-3-p yr id in yl)-2-[4-(m eth yl su lfid e)p h en yl]-
eth a n on e 15. To a slurry of magnesium turnings (191 g, 7.86
mol) in toluene (4 L) was added a solution of 4-thiomethyl-
benzyl chloride (566 g, 3.28 mol) in THF (0.545 L, 6.73 mol),
maintaining the temperature below 36 °C. The Grignard
solution is transferred to a solution of Weinreb amide (300 g,
1.66 mol) in toluene (1.7 L), maintaining the temperature
below - 10 °C. The mixture was aged for 1 h and then
quenched by the addition of aqueous acetic acid (50 wt %, 0.5
L). Toluene and water were added and the layers separated.
The aqueous layer was extracted with toluene (×2) and
concentrated to a yellow solid. Recrystallization from isopropyl
acetate/heptane (9 L) gave the ketosulfide as a colorless solid
1
(342 g) in 80% isolated yield: mp 111-113 °C; H NMR (400
MHz CDCl3) δ 9.09 (d, 1H, J ) 1.6 Hz), 8.12 (dd, 1H, J 1)8.1
J O000870Z