Synthesis of a substance P antagonist: An efficient synthesis of 5-substituted-4-N,N-dimethylamino-1,2,3-triazoles

Article abstract of DOI:10.1021/op050019s

A highly efficient synthesis of the substance P antagonist 1 is reported starting from the optically pure morpholine acetal derivative 2. The 5-dimethylaminomethyl-1,2,3-triazole moiety is elaborated via a 1,3-dipolar cycloaddition between an activated acetylenic intermediate and sodium azide. Two approaches to the construction of the triazole moiety of 1 have been designed. The first approach is linear affording the side chain in four steps and 85-92% overall yield. The reactive acetylenic aldehyde 5 allowed for a mild azide cyclization. A simple reductive amination of the triazole aldehyde completed the synthesis of 1. An alternative, more efficient, convergent synthesis using analogous methodology developed from the linear synthesis was designed to improve the overall efficiency of the process as well as remove the concerns with the handling of azide. The triazole adduct was prepared offline in a two-step, one-pot operation as the building block 4-N,N-dimethyl-aminomethyl-1, 2,3-triazole-aldehyde 3. Formylation of N,N-dimethylaminopropyne and azide cyclization were carried out as a through process to afford the triazole aldehyde 3 in 90% assay yield. The product was isolated in overall yields ranging from 67 to 81% depending on method. The aldehyde group of 3 was used for coupling to the morpholine building block through a reductive amination with NaBH(OAc)₃ in near quantitative yield to afford the substance P antagonist 1 as a hydrochloride salt in 95% yield from 2.

Full text of DOI:10.1021/op050019s

Organic Process Research & Development 2005, 9, 490−498
Synthesis of a Substance P Antagonist: An Efficient Synthesis of
5-Substituted-4-N,N-dimethylamino-1,2,3-triazoles
Michel Journet,* Dongwei Cai,* David L. Hughes, Jason J. Kowal, Robert D. Larsen,‡ and Paul J. Reider‡
Department of Process Research, Merck Research Laboratories, P.O. Box 2000, RY800-B267, Rahway,
New Jersey 07065-0900, U.S.A.
Scheme 1
Abstract:
A highly efficient synthesis of the substance P antagonist 1 is
reported starting from the optically pure morpholine acetal
derivative 2. The 5-dimethylaminomethyl-1,2,3-triazole moiety
is elaborated via a 1,3-dipolar cycloaddition between an
activated acetylenic intermediate and sodium azide. Two ap-
proaches to the construction of the triazole moiety of 1 have
been designed. The first approach is linear affording the side
chain in four steps and 85-92% overall yield. The reactive
acetylenic aldehyde 5 allowed for a mild azide cyclization. A
simple reductive amination of the triazole aldehyde completed
the synthesis of 1. An alternative, more efficient, convergent
synthesis using analogous methodology developed from the
linear synthesis was designed to improve the overall efficiency
of the process as well as remove the concerns with the handling
of azide. The triazole adduct was prepared offline in a two-
step, one-pot operation as the building block 4-N,N-dimethyl-
aminomethyl-1,2,3-triazole-aldehyde 3. Formylation of N,N-
dimethylaminopropyne and azide cyclization were carried out
as a through process to afford the triazole aldehyde 3 in 90%
assay yield. The product was isolated in overall yields ranging
from 67 to 81% depending on method. The aldehyde group of
3 was used for coupling to the morpholine building block
through a reductive amination with NaBH(OAc)3 in near
quantitative yield to afford the substance P antagonist 1 as a
hydrochloride salt in 95% yield from 2.
compound 1 as a water-soluble3d hydrochloride salt with
desirable physical properties. Herein, we report two efficient
syntheses of this substance P antagonist 1 starting from the
optically pure morpholine derivative 22 (Scheme 1). In both
routes the triazole is produced through a 1,3-dipolar cy-
cloaddition of an acetylenic aldehyde with sodium azide. In
the first approach the dimethylaminomethyltriazole is con-
structed upon the morpholine in four linear steps,5 whereas
in the second it is prepared in two steps offline from
(2) Brands, K. M. J.; Payack, J. F.; Rosen, J. D.; Nelson, T. D.; Candelario,
A.; Huffman, M. A.; Zhao, M. M.; Li, J.; Craig, B.; Song, Z. J.; Tschaen,
D. M.; Hansen, K.; Devine, P. N.; Pye, P. J.; Rossen, K.; Dormer, P. G.;
Reamer, R. A.; Welch, C. J.; Mathre, D. J.; Tsou, N. N.; McNamara, J. M.;
Reider, P. J. J. Am. Chem. Soc. 2003, 125, 2129-2135 and references
therein.
Introduction
Substance P is a peptide composed of 11 amino acids
which binds to the neurokinin receptor (NK-1). Substance
P (NK-1) receptor antagonists1 have been shown to be
potential therapeutic agents for a wide variety of important
medical disorders. Among these was found a substance P
antagonist that is effective in the treatment of chemotherapy-
induced nausea and vomiting which led Merck to the
identification of Aprepitant.2 Heterocyclic-substituted meth-
ylmorpholine acetals have been identified as potent, orally
bioavailable NK-1 receptor antagonists.3,4 Substitution with
a 5-dimethylaminomethyl-1,2,3-triazole4 resulted in potent
(3) (a) Kramer, M. S.; Cutler, N.; Feighner, J.; Shrivastava, R.; Carman, J.;
Sramek, J. J.; Reines, S. A.; Guanghan, L.; Snavely, D.; Wyatt-Knowles,
E.; Hale, J. J.; Mills, S. G.; MacCoss, M.; Swain, C. J.; Harrison, T.; Hill,
R. A.; Hefti, F.; Scolnick, E. M.; Cascieri, M. A.; Chicchi, G. G.; Sadowski,
S.; Williams, A. R.; Hewson, L.; Smith, D.; Carlson, E. J.; Hargreaves, R.
J.; Rupniak, N. M. J. Science 1998, 281, 1640. (b) Rupniak, N. M. J.;
Kramer, M. S. Trends Pharmacol. Sci. 1999, 20, 485. (c) Hale, J. J.; Mills,
S. G.; MacCoss, M.; Shah, S. K.; Qi, H.; Mathre, D. J.; Cascieri, M. A.;
Sadowski, C. D.; MacIntyre, D. E.; Metzger, J. M. J. Med. Chem. 1996,
39, 1760. (d) Hale, J. J.; Mill, S. G.; MacCoss, M.; Finke, P. E.; Cascieri,
M. A.; Sadowski, S.; Ber, E.; Chicchi, G. G.; Kurtz, M.; Metzger, J.;
Eiermann, G.; Tsou, N. N.; Tattersall, D.; Rupniak, N. M. J.; Williams, A.
R.; Rycroft, W.; Hargreaves, R.; MacIntyre, D. E. J. Med. Chem. 1998, 41,
4607. (e) Hale, J. J.; Mills, S. G.; MacCoss, M.; Dorn, C. P.; Finke, P. E.;
Budhu, R. J.; Reamer, R. A.; Huskey, S.-E. W.; Luffer-Atlas, D.; Dean. B.
J.; McGowan, E. M.; Feeney, W. P.; Chiu, S.-H. L.; Cascieri, M. A.; Chicchi,
G. G.; Kurtz, M. M.; Sadowski, S.; Ber, E.; Tattersall, D.; Rupniak, N. M.
J.; Williams, A. R.; Rycroft, W.; Hargreaves, R.; Metzger, J. M.; MacIntyre,
D. E. J. Med. Chem. 2000, 43, 1234.
* To whom correspondence should be addressed. E-mail: michel_journet@
merck.com and dongwei_cai@merck.com.
‡ Current address: Amgen, One Amgen, Centre Dr., Thousand Oaks, CA
91320.
(1) (a) Gardner, C. J.; Twissell, D. J.; Dale, T. J.; Gale, J. D.; Jordan, C. C.;
Kilpatrick, G. J.; Bountra, C.; Ward, P. Br. J. Pharmacol. 1995, 116, 3158.
(b) Swain, C. J. Prog. Med. Chem. 1998, 35, 57. (c) Seward, E. M.; Swain,
C. J. Expert Opin. Ther. Pat. 1999, 9, 571. (d) Swain, C.; Rupniak, N. M.
J. Ann. Rep. Med. Chem. 1999, 34, 51.
(4) Harrison, T.; Owens, A. P.; Williams, B. J.; Swain, C. J.; Williams, A.;
Carlson, E. J.; Rycroft, W.; Tattersall, F. D.; Cascieri, M. A.; Chicchi, G.
G.; Sadowski, S.; Rupniak, N. M. J.; Hargreaves, R. J. J. Med. Chem. 2001,
44, 4296.
(5) Cai, D.; Journet, M.; Larsen, R. D. U.S. Patent 6,051,707, 2000.
490
•
Vol. 9, No. 4, 2005 / Organic Process Research & Development
10.1021/op050019s CCC: $30.25 © 2005 American Chemical Society
Published on Web 06/14/2005

Scheme 2 a
Scheme 4
a Reagents and conditions: (a) Propargyl bromide, K2CO3, DMF, rt, 99%;
(b) n-BuLi, -60 °C; Me2NCOCl, THF, 85%; (c) NaN3, AcOH, DMSO, 75 °C;
(d) DIBAL-H, PhCH3, 84% overall from 3.
dimethylaminopropyne and coupled subsequently in a con-
vergent manner.
Results and Discussion
Linear Approach. The first-generation synthesis of 1
from 2 used a four-step sequence to incorporate the di-
methylaminotriazole ring. Reaction of 2 with propargyl
bromide in the presence of K2CO3/DMF gave rise to 6 in
99% yield. Deprotonation with butyllithium and addition of
dimethylcarbamoyl chloride provided the propargylic amide
4. The remainder of the synthesis presented some issues for
future development. The sluggish azide cyclization of 4 to
7 required heating at 75 °C and 2 equiv of acetic acid to
avoid the formation of the Michael-addition byproducts, such
as 8.
introduction of the dimethylamino group through reductive
amination of the triazole aldehyde 9.
The success of this new approach depended on the
development of an efficient and scalable synthesis of the key
intermediate 5. No general, straightforward, high-yielding
reactions were found in the literature for the preparation of
R,â-acetylenic aldehydes from terminal alkynes. This was
actually quite surprising since such alkynals have been widely
used as Michael acceptors,6 dienophiles in Diels-Alder
reactions,7 or as key intermediates in the synthesis of
important natural products, such as cembranolide,8a calichea-
micin,8b or neocarzinostatin.9 A two-step procedure involving
addition of paraformaldehyde followed by oxidation has been
reported.8,10b-c In some isolated cases, formamide derivatives
other than DMF have been applied, increasing the yields to
80-90%.11
As reported by Olah,11b the formylating reagent does not
need to bear additional ligands to stabilize the intermediate
R-aminoalkoxide making the cheap, readily available DMF
a potentially convenient reagent. As a test of this concept,
the alkyne 6 was treated with n-BuLi followed by the
addition of DMF. In fact, the incipient intermediate aldehyde
10 (Scheme 4) was stable under the reaction conditions.
However, the moderate yields in the direct formylation of
the acetylide with DMF9,10a,12 were due to the quench of 10.
The source of the previously reported low yields has been
related with the quench where side-reactions of the product
For the reduction of the amide to the dimethylamine of 1,
an inverse addition of the amide into a DIBAL-H solution
was essential for achieving a high conversion. To separate
the product in high recovery, a quench with saturated sodium
potassium tartrate was necessary. This process to 1 was
carried out in a moderate 60% overall isolated yield from
the secondary amine 2 (Scheme 2).
A second-generation process5 was designed to overcome
the hazardous triazole formation and to increase the overall
yield of the sequence (Scheme 3). The ease of an azide-
alkyne cyclization to a triazole is governed by the polariza-
tion of the acetylene. This led us to consider alternative
functional groups. Potentially, the intermediacy of the R,â-
acetylenic aldehyde 5 could afford a more reactive system
towards azide addition as well as providing a handle for
(6) Covarrubias-Zu´n˜iga, A.; Rios-Barrios, E. J. Org. Chem. 1997, 62, 5688.
(7) (a) Corey, E. J.; Lee, T. W. Tetrahedron Lett. 1997, 38, 5755. (b) Ishihara,
K.; Kondo, S.; Kurihara, H.; Yamamoto, H. J. Org. Chem. 1997, 62, 3026.
(8) (a) Marshall, J. A.; Crooks, S. L.; DeHoff, B. S. J. Org. Chem. 1988, 53,
1616-1623. (b) Marshall, J. A.; Andersen, M. W. J. Org. Chem. 1992, 57,
2766-2768.
(9) Mikami, K.; Matsueda, H.; Nakai, T. Tetrahedron Lett. 1993, 34, 3571-
3572.
Scheme 3
(10) (a) Mikami, K.; Matsueda, H.; Nakai, T. Synlett 1993, 23-25. (b) Petassis,
N. A.; Teets, K. A. Tetrahedron Lett. 1993, 34, 805-808. (c) Rucker, M.;
Bru¨ckner, R. Tetrahedron Lett. 1997, 38, 7353.
(11) (a) For instance, using N-formylmorpholine, 3-phenylpropynal was syn-
thesized from lithium phenyl acetylide in 80% yield: Olah, G. A.;
Ohannesian, L.; Arvanaghi, M. J. Org. Chem. 1984, 49, 3856-3857. (b)
Olah, G. A.; Arvanaghi, M. Angew. Chem., Int. Ed. Engl. 1981, 20, 878.
(c) Comins, D.; Meyers, A. I. Synthesis 1978, 403. (d) Amaratunga, W.;
Fre´chet, J. M. J. Tetrahedron Lett. 1983, 24, 1143-1146.
(12) Jones, E. R. H.; Skattebol, L.; Whiting, M. C. J. Chem. Soc. 1958, 1054-
1059.
Vol. 9, No. 4, 2005 / Organic Process Research & Development
•
491

Scheme 5
Scheme 6
reaction, free of this undesired byproduct, a precise titration15
and addition of n-butyllithium would be required.
An alternative base was sought to avoid the over-
deprotonation. Interestingly, even the bulky LDA generated
0.4 area % of the impurity with 15 mol % excess of base.
Ethylmagnesium chloride proved to be the base of choice.
The reaction could be run at room temperature. Even with a
20 mol % excess of the base, 14 was not detected.
can predominate. In the quench the intermediate 10 collapses
to release the strong nucleophile dimethylamine, which can
subsequently react with the alkynal 5. Three byproducts 11-
13 were isolated from the formylation. Their formation
depended on the way the reaction was hydrolyzed. In the
quench with water, 5 was isolated in <50% yield, 12 being
the major byproduct. In an acidic workup with aqueous acetic
acid the stable iminium acetate 11 was formed, which was
slowly hydrolyzed to the enaminoketone 13 upon aging.
Unfortunately, the workup needed to be re-designed.
Previously, the phosphate buffer was used to prevent the
generated dimethylamine from adding to the acetylenic
aldehyde.16 However, with the Grignard reagent, insoluble
magnesium salts formed. Since the quench needed to be run
at a pH between 4 and 6 to avoid the formation of the
Michael adducts, several alternative buffers were screened.
The monosodium salt of citric acid offered a buffer zone
similar to that by KH2PO4. Since the substrate 6 was isolated
as a toluene solution, this solution was used directly in the
formylation. EtMgCl (1.2 equiv) was added at room tem-
perature followed by addition of DMF (2 equiv). The reverse-
quench of 10 into a biphasic solution of 7% aqueous sodium
dihydrogen citrate (2 equiv, pH ≈ 3.9-4.5) and toluene
produced two clear phases. Acetylenic aldehyde 5 was
completely stable under these conditions and was synthesized
in 99% yield. The only impurities were the remaining starting
material 6 (0.35 area %) and the iminium Michael adduct
13 (0.25 area %), which was extracted into the aqueous layer.
Formation of 5 was so clean that the crude product was
carried through to the triazole formation (Scheme 3) without
purification. Fortuitously, the toluene solution of 5 could be
used directly in the triazole formation after azeotropic
distillation (KF < 250 µg water/mL) again obviating a
solvent switch. The azide cyclization as in the original
synthesis was best performed in DMSO. All attempts to run
this reaction in another solvent met with failure. Interestingly,
only 5% of the triazole aldehyde 9 was formed in methanol
where the major impurity (25%) was the isoxazole 16
(Scheme 6; R same as in Scheme 4). This indicates that the
1,3-dipolar cycloaddition does not proceed by a concerted
mechanism but more likely through a conjugate addition of
the azide followed by cyclization. Unlike the acetylenic
amide 4, the cyclization of 5 occurred instantaneously and
quantitatively in DMSO at room temperature to give the
penultimate triazole aldehyde 9. With the aldehyde, the
reactivity was so high that acetic acid was unnecessary. This
was a major accomplishment in that it avoided the potential
generation of HN3, a highly explosive gas.17 As previously
observed, good mixing was tremendously important in this
1
When harsher conditions were used (3 N HCl), crude H
NMR showed only the acetylenic aldehyde 5, but the isolated
yield was only 40-45%. In fact, reverse-phase HPLC
analysis13 of the aqueous layer revealed that the three Michael
adducts 11, 12, and 13 had formed and partitioned into the
aqueous layer as their hydrochloride salts. These three
byproducts were eventually isolated and fully characterized
by NMR to confirm their structure. Their formation was
avoided through the inverse addition of the R-aminoalkoxide
mixture into a buffered biphasic mixture of 10% aqueous
monobasic potassium phosphate (4 equiv) and MTBE.
Thereby, the product 5 was suitably stabilized from over-
addition of the reaction components. The reaction itself
proceeded in nearly quantitative yield. Because of the
stablized R-aminoalkoxide intermediate and inverse quench,
no over-addition from butyllithium was observed.
With this method, the terminal alkyne 6 was formylated
under the optimized conditions to afford 5 in 97% assay yield
with only 1 area % of the alkyne remaining. The iminium
Michael adduct 11 was generated in only 0.25 area %
(detected at λ ) 325 nm) and was completely removed in
the aqueous layer upon workup. The enamine 12 and
ketoenamine 13 were generated in <0.1 area %.
A problem in this reaction was that only one equivalent
of n-butyllithium could be tolerated; otherwise, even with a
trace excess, abstraction of the aromatic proton adjacent to
the trifluoromethyl group occurred to give the aldehyde
byproduct 14, which was converted into the hydroxymethyl
derivative 15 in the final step of the process (Scheme 5).14
The impurity 14 was generated in ∼0.2 area % when an
excess of 2-5 mol % of n-butyllithium was used. A stress
reaction using 100 mol % excess of n-butyllithium resulted
in the formation of 10 area % of the aldehyde 14 confirming
that the impurity was generated from the overcharge of base.
The eventual new process impurity 15 was poorly rejected
upon crystallization of the HCl salt of 1. To get a clean
(13) All Michael adducts have a strong absorption at 325 nm.
(14) Under the conditions we used for the reductive amination of triazole aldehyde
9 with borane‚dimethylamine, aldehyde 14 was reduced to the corresponding
hydroxymethyl derivative 15.
(15) Suffert, J. J. Org. Chem. 1989, 54, 509-510.
(16) Journet, M.; Cai, D.; DiMichele, L. M.; Larsen, R. D. Tetrahedron Lett.
1998, 39, 6427-6428.
(17) Stohlmeier, M.; Thewalt, K. Chem. Ind. Dig. 1993, 6, 124-128.
492
•
Vol. 9, No. 4, 2005 / Organic Process Research & Development

Scheme 7
Scheme 9
reaction since any starvation of azide during the addition
led to some decomposition.18 With this new procedure the
triazole aldehyde 9 was prepared in 99% assay yield (Scheme
3).
At this stage of the synthesis, the penultimate intermediate
9 had been synthesized in 97% assay yield starting from the
secondary amine 2. To develop a through process and obviate
the workup between the reactions, the reductive amination
was investigated in DMSO. Fortunately, DMSO was the best
solvent for the reduction. Only ∼0.1% alcohol byproduct
was generated as compared to ∼1% alcohol in other solvents,
such as methanol, ethanol, 2-propanol, or THF. A 40 wt %
aqueous solution of dimethylamine was the most convenient
source of the amine. The readily available borane‚dimeth-
ylamine complex proved to be a highly effective reducing
agent. A more reactive borane‚amine complex, such as
borane‚pyridine converted to borane‚dimethylamine in situ,
whereas a less reactive borane‚amine, such as borane‚tri-
ethylamine required forcing conditions. As a throughput
process with no workup, 40 wt % aqueous dimethylamine
was added to the DMSO/toluene mixture of 9, followed with
acetic acid and borane‚dimethylamine. The reaction mixture
was aged at 40 °C for 12-18 h, and the product 1 was
generated in 98% assay yield (Scheme 3). The final product
1 was crystallized as the hydrochloride salt (4 N HCl in
isopropyl alcohol) from toluene/heptane in 89% overall yield
from 2.
Convergent Approach. Although the linear process was
very efficient and high yielding (89-92% overall), none of
the intermediates was crystalline. This necessitated the
development of a throughput process that, although efficient,
lacked flexibility in the preparation of this compound. In
addition there were concerns with running the triazole
preparation with excess azide and the addition of acetic acid
subsequently for the reductive amination. Rather than
construct the triazole on the morpholine scaffold in a linear
approach (Scheme 3), a process where the triazole was
constructed offline and added directly in the last step would
afford an efficient, convergent synthesis of 1 (Scheme 7).19
Two approaches for coupling a triazole building block to 3
were considered. A logical strategy could employ a direct
alkylation of the secondary amine 3 with a halomethyl-1,2,3-
triazole. In contrast, a triazole carboxaldehyde derivative
could be used in a reductive amination. Both approaches were
considered as is disclosed herein.
In the alkylation approach the bromomethyl triazole 17
was prepared easily from THP-protected propargyl alcohol
(Scheme 8).20 Alkylation of the morpholine 2 effectively
intercepted the synthesis of 1 at compound 9 (Scheme 3).
An attempt to design a more convergent route through the
synthesis of 18 did indeed give 1 in one step from 2, but the
over-alkylation to 19 could not be controlled (Scheme 9).
The reductive amination approach to 1 was tested with
the aldehyde 20 as prepared in Scheme 9. To our delight,
the reductive amination of 2 with 20 using borane‚
triethylamine as the reducing agent in ethanol gave the
coupled product 21 in 95% conversion (Scheme 10). With
the feasibility of this approach confirmed, the synthesis of
the optimal coupling intermediate 3 was designed. Since the
developed synthesis of the dimethylaminomethyltriazole
group of 1 was incredibly efficient and high-yielding, a
similar strategy to the preparation 9 (Scheme 3) was followed
through formylation of N,N-dimethylaminopropyne (22) and
1,3-dipolar cycloaddition of the alkyne with sodium azide
(Scheme 11).
Commercially available 1-(dimethylamino)-2-propyne
(22) was chosen as the starting material (Scheme 11). The
substrate was deprotonated and formylated as previously
developed. Due to the greater water solubility of the
dimethylamino-substituted intermediates, as opposed to the
protected propargyl alcohols in Schemes 8 and 9, the workup
became a challenge. Moreover the aldehyde 24 was unstable
above -30 °C even under anhydrous conditions due to
polymerization and could only survive this temperature for
a couple of hours. As a result, the biphasic quench with
aqueous sodium dihydrogen citrate/toluene and the stepwise
sequence developed in the linear process for 1 could not be
used.
Scheme 10
Scheme 8
Vol. 9, No. 4, 2005 / Organic Process Research & Development
•
493

Scheme 11
Scheme 12
had to be avoided. Therefore, the process was conducted with
0.95 equiv of sodium azide affording 3 in 90% assay yield
from dimethylaminopropyne (95% from azide). This yield
penalty was to be expected since with a deficit of azide the
unreacted 24 polymerizes. An acceptable level of residual
sodium azide was reached, which was assayed at e30 ppm.21
Although a quantitative assay yield of 3 could be obtained
with 1.0 equiv of azide, while maintaining the azide level at
<30 ppm at the end of the reaction, the greater safety margin
with 0.95 equiv was preferable.
The high water solubility of 3 required changes in the
generic process to 4-substituted-1,2,3-triazole-5-carboxalde-
hyes.22 A direct extraction out of the reaction media was
not feasible with the presence of DMSO. Replacement of
DMSO was not an option since the cyclization with sodium
azide only proceeded well in this solvent. The preparation
of derivatives of the triazole, such as benzyl or trimethylsilyl,
through cyclization with alternative azide reagents failed.
Precipitation of an insoluble salt, such as calcium or barium,
or the formation of an insoluble bisulfite adduct23 was also
unsuccessful. Three different isolation processes were even-
tually developed to give 3 as a crystalline compound ranging
in 67-81% overall isolated yield from 22. Ion exchange
chromatography using Dowex resins gave the best recovery
(81% overall yield) and was used to prepare 5 kg of the
triazole aldehyde. However, the large volumes of solvent
made this isolation unsuitable for a long-term process of the
drug.
Deprotonation of the alkyne 22 was initially achieved with
n-BuLi. The isolation of 24 was not possible due to its
instability, requiring the quench and age to be carried out at
low temperatures. The intermediate lithium alkoxide 23 was
hydrolyzed with anhydrous TFA at -30 °C ,and the resulting
acetylenic aldehyde 24 was added into a DMSO solution of
sodium azide at room temperature to give 3 in a moderate
50% assay yield. Some byproducts were identified as
Michael adducts resulting from the addition of dimethylamine
onto the very reactive electrophile 24, as observed by us
previously.16 Perhaps a less nucleophilic amine could be
generated by changing the formylating reagent. Indeed,
N-methylformanilide, generating the less reactive byproduct
N-methylaniline, increased the assay yield to 90% repro-
ducibly.
To overcome the special handling required with the
reactive intermediate 24 the direct quench of 23b into sodium
azide was investigated. In our study of the synthesis of
variously substituted acetylenic aldehydes,20 the magnesium
or lithium alkoxide intermediates proved to be stable at room
temperature for at least 24 h under anhydrous conditions.
This was also true in the case of the formylation of
1-(dimethylamino)-2-propyne. However, a fast hydrolysis
occurred upon addition of water to give the corresponding
aldehyde 24, which polymerized readily to give black tars.
On the basis of these simple observations, we were confident
that the stable magnesium aminoalkoxide intermediate 23b
could be added into a wet DMSO solution of sodium azide
to give the triazole aldehyde 3 directly (Scheme 12).
1-Dimethylamino-2-propyne was deprotonated at room tem-
perature with 1.1 equiv of ethylmagnesium chloride in THF
and was reacted with 1.2 equiv of N-methylformanilide. The
resulting magnesium aminoalkoxide 23b was aged at room
temperature for an hour and added at room temperature into
a vigorously stirred 0.5 M DMSO solution of 1.1 equiv of
sodium azide containing 1.5 equiv of water. Good mixing
was tremendously important since any starvation of azide
led to partial decomposition of the liberated 24. This
throughput process using only two vessels afforded the
triazole aldehyde in quantitative assay yield.
Interestingly, we found that the piperidine adduct 25 could
be crystallized directly from the reaction mixture (DMSO).
Depending on the size of the ring, the resulting adduct
had two different structures (Scheme 13). Cyclic secondary
At the end of the reaction there was still an excess of
sodium azide remaining. To address the safety concerns in
quenching the reaction mixture with an acid the residual azide
(21) Sodium azide was assayed by reverse phase HPLC using a C-18 Metachem
inertsil ODS-3 column (250 mm × 4.6 mm, 5 µm) with a flow rate of 0.75
mL/min and using UV detection at 200 nm. Elution was a gradient using a
10/90 mixture of acetonitrile/water (0.1% H3PO4) to 30/70 in 20 min.
Retention time for azide was 7.35 min.
(22) Triazole aldehyde 3 is highly water soluble. Its solubility is 63 g/L in
methanol and DMSO, 19 g/L in DMF and <4 g/L in 2-propanol, 2-butanol,
chloroform, THF, and toluene.
(18) In the absence of sodium azide, the anion of triazole aldehyde 8 may react
with acetylenic aldehyde 4 as observed with 7.
(19) Cai, D.; Journet, M.; Kowal, J.; Larsen, R. D. U.S. Patent 6,051,717, 2000.
(20) Journet, M.; Cai, D.; Kowal, J. J.; Larsen, R. D. Tetrahedron Lett. 2001,
42, 9117-9117.
(23) Young, P. R.; Jencks, W. P. J. Am. Chem. Soc. 1978, 100, 1228-1235.
494
•
Vol. 9, No. 4, 2005 / Organic Process Research & Development

Scheme 13
Scheme 15
In 2-propanol with 1 equiv of acetic acid the reaction gave
a 92% assay yield with ∼95 area % purity. Many of these
impurities were stable borane complexes, in particular, the
borane complex with the dimethylamine moiety. To over-
come the formation of these borane‚amine complexes,
sodium triacetoxyborohydride (STAB) was tested.24 Acetic
acid was not needed with STAB. In alcohol solvents the
conversion was not good. THF worked quite well, but 2
equiv of aldehyde was needed to drive the reaction to
completion due to the limited solubility of 3 in this system.
Although DMSO was a suitable medium for the reductive
amination, the reaction intermediates were not soluble,
forming a gummy precipitate. In DMF the reaction worked
very well with >99% conversion even with 1.1 equiv of 3.
When the secondary amine p-toluenesulfonic acid salt was
treated with STAB in DMF, the N-formyl analogue of 2 was
generated at ∼0.1%/h at room temperature. Without STAB,
no formylation occurred. STAB or some impurities in the
STAB catalyzed this formylation with DMF which was
eventually replaced with the more stable N,N-dimethyl-
acetamide (DMAC). With DMAC no corresponding acety-
lation of the secondary amine occurred, and the reductive
amination performed essentially the same as in DMF. With
DMAC as solvent, the reaction performed perfectly in
quantitative assay yield, generating product of >99% purity
with only 0.1% of the secondary amine 2 remaining. This
reaction was demonstrated on 4.4 mol scale in 95% isolated
yield (99.6% reaction yield). Overall the process from the
triazole aldehyde, prepared as throughput process from
dimethylaminopropyne, was highly effective for preparing
1 (Scheme 16).
amines other than a six-membered ring gave imminium
adducts. However, when 3 was reacted with piperidine, a
white crystalline product was produced in high recovery
(75% isolated yield) with a good purity. It turned out that in
this case the imminium was not observed but formation of
the dimer adduct 25 occurred instead. The difference of
solubility in DMSO of all of these adducts was huge: 25
was essentially insoluble (1 g/L) versus 58 and 35 g/L for
the imminium‚pyrrolidine adduct 29 and hexamethylene‚
imine adduct 30, respectively. Other six-membered ring
secondary amines were tested, and all of them gave a dimer
adduct 26-28.
The next step was the cleavage of 25 into 3 that could be
achieved under acidic conditions in the presence of at least
0.5 equiv of water. The solvent of choice for this transforma-
tion was a 98/2 mixture of 2-propanol/water (7 mL/g).
Several acids were tried. Acetic acid was inefficient, whereas
the use of 2 equiv of TFA (Scheme 14) or methanesulfonic
acid or trichloroacetic acetic acid all led to the crystallized
heterocycle in 84% isolated yield. TFA was the best choice
since 1H NMR showed a very clean product with no
contamination whatsoever of piperidine. Indeed, piperidine
byproduct was completely removed as the TFA salt in the
mother liquors, whereas trichloroacetic acetic led to a
substantial amount of piperidine in the final isolated material.
An isolated yield of 84% was the maximun we could get
since the solubility of pure 3 in a 98/2 mixture of 2-propanol/
water (10 mL/g) in the presence of one equiv of TFA salt of
piperidine was 16 g/L.
Scheme 16 a
Scheme 14
a Reagents and conditions: (a) (i) EtMgCl, THF, rt; (ii) PhN(Me)CHO, rt;
(iii) NaN3, DMSO, rt; (b) 2, NaBH(OAc)3, DMAC, 0-5 ° C.
Conclusion
The preferred isolation process utilized a reverse extrac-
tion of the sodium salt of 3 into toluene with Aliquat 336
(Scheme 15). The triazole was then extracted into aqueous
acetic acid. Removal of the water and acetic acid by
evaporative displacement with 2-butanol gave crystalline
triazole aldehyde 3 in 67% overall yield.
The convergent synthesis was completed by reductive
amination of 2 with 3. As before borane‚triethylamine was
tested as the reducing agent in a solvent mixture of
2-propanol/DMSO containing acetic acid. Although the
desired product 1 was obtained, the reaction was not clean.
Two processes for the synthesis of 1 have been developed.
An efficient four-step linear process to prepare the substance
P antagonist in 85-92% overall yield from the key inter-
mediate 2 has been described. This process has been fully
demonstrated with the preparation of kilogram quantities of
1. To further improve upon the chemistry developed for the
construction of the dimethylaminomethyltriazole a modifica-
tion of the process was carried out offline to afford a
convergent approach to 1. A practical synthesis and isolation
of triazole aldehyde 2 has been developed. A reductive
amination strategy to join the triazole aldehyde 3 and the
Vol. 9, No. 4, 2005 / Organic Process Research & Development
•
495

secondary amine 2 using sodium triacetoxyborohydride
afforded a high yielding, efficient preparation of the sub-
stance P antagonist in >95% yield.
2-(R)-(1-(R)-(3,5-Bis(trifluoromethyl)phenyl)ethoxy)-3-
(S)-4-(4-oxo-but-2-ynyl-(4-fluorophenyl))morpholine (5).
To a stirred solution of crude 2-(R)-(1-(R)-(3,5-bis(trifluo-
romethyl)phenyl)ethoxy)-3-(S)-(4-fluorophenyl)-4-propargyl-
morpholine (6) (470.5 g assay, 0.99 mol) in dry toluene (3.4
L) was added ethylmagnesium chloride (2.0 M/THF, 595.0
mL, 1.19 mol) over 20 min, maintaining the temperature
between 15 and 25 °C. The resulting solution was stirred
for 2 h at room temperature, and anhydrous DMF (KF e
100 µg H2O/mL, 153.0 mL, 1.98 mol) was added over 10
min, maintaining the temperature between 20 and 25 °C.
The solution was aged for an additional hour and was reverse
added into a vigorously stirred biphasic solution prepared
from sodium dihydrogen citrate (424.0 g, 1.98 mol) in water
(5.6 L, ∼7 wt %) and toluene (5.6 L) at room temperature
(final pH∼3.9 to 4.2). The biphasic reaction mixture was
vigorously stirred for 4 h at room temperature, and the layers
were separated. The organic layer was washed with water
(2 × 4.0 L) and was concentrated under reduced pressure to
give 5 as a yellow oil (493.5 g assay) in 99% yield, which
was used as is in the next step. A standard was purified as
an oil by flash chromatography on silica gel eluting with a
80:20 mixture of hexane/ethyl acetate (Rf 0.3). [R]D25 +225
Experimental Section
General. Reactions were performed under a positive
atmosphere of dry nitrogen. Prior to use, the solvents were
dried over 4 Å molecular sieves to <100 µg H2O/mL. Water
content was determined by Karl Fisher titration (KF).
Commercially available reagents were purchased from Al-
drich Chemical Co. and used as received. Melting points are
uncorrected. Elemental analyses were performed by Quan-
titative Technologies Inc., Whitehouse, NJ.
2-(R)-(1-(R)-(3,5-Bis(trifluoromethyl)phenyl)ethoxy)-3-
(S)-(4-fluorophenyl)-4-propargylmorpholine (6). To a
stirred solution of 2-(R)-(1-(R)-(3,5-bis(trifluoromethyl)-
phenyl)ethoxy)-3-(S)-(4-fluorophenyl)morpholine p-toluene-
sulfonate salt (2) (609.0 g, 1.0 mol) in DMF (3.4 L) was
added powdered potassium carbonate 325 mesh (346.0 g,
2.5 mol) at room temperature in one portion followed by
the addition of propargyl bromide (80 wt % in toluene, 134.0
mL, 1.2 mol) over 5 min. The reaction was sligthly
exothermic reaching ∼29 °C. The reaction mixture was
stirred at room temperature for 3.5 h (>99.9% conversion)
and aqueous dimethylamine (40 wt %, 38.0 mL, 0.30 mol)
was added in one portion and aged 0.5 h to convert the excess
of propargyl bromide into the less toxic and water soluble
1-(dimethylamino)-2-propyne. The reaction mixture was
diluted with toluene (6.7 L) and water (5.4 L), and the layers
were separated. The organic layer was washed with water
(2 × 3.4 L) and was finally concentrated under reduced
pressure to give 6 as a yellow oil (470.5 g assay) in 99%
yield, which was used as is in the next step. A standard as
an oil was purified by flash chromatography on silica gel
eluting with an 85:15 mixture of hexane/ethyl acetate (Rf
1
(c 1.0, CHCl3). H NMR (400 MHz, CD3CN) δ 9.15 (s, 1
H), 7.75 (s, 1 H), 7.41 (br t, J ) 6.8 Hz, 2 H), 7.33 (s, 2 H),
7.05 (t, J ) 8.6 Hz, 2 H), 4.89 (q, J ) 6.6 Hz, 1 H), 4.37 (d,
J ) 2.8 Hz, 1 H), 4.26 (td, J ) 11.5 and 2.8 Hz, 1 H), 3.67
(ddd, J ) 11.4, 3.2, and 1.6 Hz, 1 H), 3.59 (d, J ) 2.8 Hz,
1 H), 3.42 (AB q, J ) 20.0 Hz, 2 H), 2.89-2.94 (m, 1 H),
2.81 (td, J ) 11.5 and 3.5 Hz, 1 H), 1.43 (d, J ) 6.4 Hz, 3
H); 13C NMR (100 MHz, CD3CN) δ 177.4, 162.5 (d, J )
244.3 Hz, Csp2-F), 146.3, 132.5, 131.4, 130.9 (q, J ) 33.1
Hz, Csp2-CF3), 126.8, 123.4 (q, J ) 272.1 Hz, CF3), 121.3,
114.8 (d, J ) 24.6 Hz), 95.5, 90.6, 85.9, 72.2, 66.5, 58.9,
51.2, 43.6, 23.4. Anal. Calcd for C24H20F7NO3 (503.41) C,
57.26; H, 4.00; N, 2.78. Found: C, 57.45; H, 4.05; N, 2.56.
2-(R)-(1-(R)-(3,5-Bis(trifluoromethyl)phenyl)ethoxy)-
4-(5-oxomethyl-1,2,3-triazol-4-yl)methyl-3-(S)-(4-fluoro-
phenyl)morpholine (9). To a vigorously stirred solution of
sodium azide (76.1 g, 1.17 mol) in DMSO (3.1 L) was added
a toluene (1.4 L) solution of crude 2-(R)-(1-(R)-(3,5-bis-
(trifluoromethyl)phenyl)ethoxy)-3-(S)-4-(4-oxo-but-2-ynyl-
(4-fluorophenyl))morpholine (5) (493.5 g, 0.98 mol) over
20 min, maintaining the temperature between 20 and 25 °C.
The resulting reaction mixture was stirred at room temper-
ature for 0.5 h to give the triazole aldehyde derivative 9
(530.0 g assay) in 99% yield. No workup was performed;
rather, the whole DMSO crude solution was carried through
to the next step as is. A standard was purified as a foamy
solid by flash chromatography on silica gel eluting with a
40:60 mixture of hexane/ethyl acetate (Rf 0.3). [R]D25 +100
25
1
0.3). [R]D +178 (c 0.58, CHCl3). H NMR (400 MHz,
CDCl3) δ 7.62 (s, 1 H), 7.39 (br t, J ) 6.4 Hz, 2 H), 7.17 (s,
2 H), 7.04 (t, J ) 8.8 Hz, 2 H), 4.90 (q, J ) 6.6 Hz, 1 H),
4.33 (d, J ) 2.8 Hz, 2 H), 3.71 (ddd, J ) 11.4, 3.2, and 1.6
Hz, 1 H), 3.62 (d, J ) 3.2 Hz, 1 H), 3.26 (d, J ) 2.7 Hz, 2
H), 2.98 (td, J ) 11.4 and 3.2 Hz, 1 H), 2.79-2.87 (m, 1
H), 2.20 (t, J ) 2.7 Hz, 1H), 1.46 (d, J ) 6.4 Hz, 3 H); 13
C
NMR (100 MHz, CDCl3) δ 162.7 (d, J ) 247.2 Hz, Csp2-
F), 145.5, 131.6 (q, J ) 33.3 Hz, Csp2-CF3), 131.0 (d, J )
8.1 Hz), 126.3, 123.0 (q, J ) 273.0 Hz, CF3), 121.4, 115.1
(d, J ) 21.1 Hz), 95.6, 74.3, 72.4, 66.4, 59.2, 51.0, 43.8,
24.4. Anal. Calcd for C23H20F7NO2 (475.40) C, 58.11; H,
4.24; N, 2.95. Found: C, 57.99; H, 4.21; N, 2.84.
(24) With various lots of purchased STAB two impurities were detected and
identified as borane complexes of 1 previously observed with borane
reducing agents. The STAB was contaminated with NaBH4 as confirmed
with solid-state boron-11 NMR study. Solution assays did not detect NaBH4
due to the rapid reaction with STAB generating sodium mono- and bis-
acetoxyborohydride. STAB was originally charged as a solid. By predis-
solving the reagent in DMAC the NaBH4 contaminant was converted to
the mixed forms of acetoxyborohydrides. As a result the borane byproducts
were no longer observed. As an alternative STAB can be prepared in situ.
NaBH4 is very soluble in DMAC. Upon mixing with excess acetic acid
STAB can be cleanly generated. Although adding either mixture to the other
will generate the same quality STAB, addition of a solution of NaBH4 into
a cold (0 °C) solution of acetic acid (3.3 equiv) in DMAC will control the
foaming caused by the release of hydrogen gas.
1
(c 0.5, CHCl3). H NMR (400 MHz, CD3CN) δ 10.02 (s, 1
H), 7.75 (s, 1 H), 7.48 (br t, J ) 7.8 Hz, 2 H), 7.32 (s, 2 H),
7.03 (t, J ) 8.9 Hz, 2 H), 4.89 (q, J ) 6.6 Hz, 1 H), 4.37 (d,
J ) 2.9 Hz, 1 H), 4.23 (br t, J ) 11.8 Hz, 1 H), 3.83 (d, J
) 15.3 Hz, 1 H), 3.58 (ddd, J ) 11.5, 3.2, and 1.6 Hz, 1 H),
3.54 (d, J ) 2.8 Hz, 1 H), 3.53 (d, J ) 15.3 Hz, 1 H), 2.83
(br d, J ) 11.6 Hz, 1 H), 2.51 (td, J ) 11.8 and 3.5 Hz, 1
H), 1.43 (d, J ) 6.4 Hz, 3 H); 13C NMR (100 MHz, CD3-
496
•
Vol. 9, No. 4, 2005 / Organic Process Research & Development

CN) δ 185.3, 162.5 (d, J ) 244.0 Hz, Csp2-F), 146.3, 143.1,
133.3, 131.4, 131.2, 130.9 (q, J ) 33.1 Hz, Csp2-CF3),
126.8, 123.4 (q, J ) 272.0 Hz, CF3), 121.2, 114.7 (d, J )
21.6 Hz), 95.6, 72.2, 68.8, 59.1, 52.4, 48.3, 23.5. Anal. Calcd
for C24H21F7N4O3 (546.44) C, 52.75; H, 3.87; N, 10.25.
Found: C, 52.85; H, 3.83; N, 10.01.
reaction mixture was aged at room temperature for 2 h.
N-Methylformanilide (162.5 g, 148.0 mL, 1.2 mol) was
added over 15 min at 20-25 °C. The reaction mixture was
stirred at room temperature for an additional hour and was
then added over 30 min into a vigorously stirred DMSO
solution (2.0 L) containing sodium azide (61.8 g, 0.95 mol)
and water (27.0 mL, 1.5 mol) at 20-25 °C to give the
triazole aldehyde (139.0 g assay) in 90% assay yield (95%
yield relative to sodium azide). No unreacted azide was
detected by assay (LOD < 30 ppm).21 4-N,N-Dimethylami-
nomethyl-5-formyl-1,2,3-triazole was isolated from the same
reaction mixture by any of the following three procedures.
Method A: Resin Column Isolation. The DMSO solution
of triazole aldehyde (pH ≈ 9.7) was pH adjusted to 7.0 with
aqueous HCl (1.0 N, ca. 2.0 L) and washed with ethyl acetate
(1.0 L). The layers were separated, and the aqueous DMSO
phase was loaded onto a column containing Dowex 50 W
X8-100 (1.7 mequiv/mL wet, 5.5 equiv, 3.3 L) at a flow
rate of 4 bed volumes per hour. The resin was then washed
with 2 bed volumes of deionized water, and the product was
eluted with a mixture of acetonitrile/water/triethylamine (6:
3:1). After 1 bed volume (3.3 L) was eluted to the column
displacing the water wash, the flow was stopped, and the
column was equilibrated for 1 h. An additional 1.5 bed
volumes (5.0 L) was eluted over an hour to provide a 97%
recovery of the triazole aldehyde. The acetonitrile/water/
triethylamine solution was concentrated under reduced pres-
sure, and the solvent was switched to 2-butanol (until water
< 0.5 vol %) to crystallize the product. The final volume
was adjusted to 750 mL. The crystalline product was filtered,
washed with 2-butanol (200 mL), and dried at 40 °C under
vacuum with a nitrogen stream to afford pure 4-N,N-
dimethylaminomethyl-5-formyl-1,2,3-triazole (125.0 g) in
81% overall yield from 1-(dimethylamino)-2-propyne and
85% overall yield from sodium azide.
Method B: Aliquat 336 Extraction Process. To the DMSO
solution of triazole aldehyde (pH ≈ 9.7) was added Aliquat
336 (807.0 g, 2.0 mol) and toluene (2.0 L). The mixture was
stirred for 0.5 h, and water was added (1.1 L). The layers
were separated, and the aqueous phase was back-extracted
twice with Aliquat 336 (200.0 g, 0.5 mol) in toluene (2.0
L). The combined organic layers were washed with water
(820 mL) to remove excess DMSO. The resulting organic
layer was washed with a 12.5 vol % of aqueous acetic acid
solution (930.0 mL, 2.0 mol) to release the triazole aldehyde
into the aqueous phase. The layers were separated, and the
organic phase was washed once again with plain deionized
water (365 mL). These combined extractions recovered the
triazole aldehyde (124.0 g) in 89% overall assay yield from
the crude DMSO solution. The aqueous solution was
concentrated under reduced pressure, and the solvent was
switched to 2-butanol (until water < 0.5 vol %) to crystallize
the product. The final volume was adjusted to 750 mL. The
crystalline product was filtered, washed with 2-butanol (200
mL), and dried at 40 °C under vacuum with a nitrogen stream
to afford pure 4-N,N-dimethylaminomethyl-5-formyl-1,2,3-
triazole (3) (102.5 g) in 67% overall yield based on
1-(dimethylamino)-2-propyne (70% relative to sodium azide).
2-(R)-(1-(R)-(3,5-Bis(trifluoromethyl)phenyl)ethoxy)-4-
(5-(dimethylamino)methyl-1,2,3-triazol-4-yl)methyl-3-(S)-
(4-fluorophenyl)morpholine Hydrochloride Salt (1). To
the crude DMSO reaction solution of 2-(R)-(1-(R)-(3,5-bis-
(trifluoromethyl)phenyl)ethoxy)-4-(5-oxomethyl-1,2,3-tri-
azol-4-yl)methyl-3-(S)-(4-fluorophenyl)morpholine (9) was
added dimethylamine (40 wt % aqueous; 1.62 L), followed
by 2-propanol (1.38 L) and acetic acid (0.55 L) while
maintaining the temperature at <40 °C. Borane‚dimethylamine
complex (59.0 g; 0.97 mol) was added in one portion, and
the reaction mixture was heated to 40 °C and aged for 16 h
with stirring. The reaction was cooled to room temperature,
diluted with water (1.5 L), and neutralized with concentrated
phosphoric acid (∼85%) to pH ≈ 8.5-9. The reaction
mixture was further diluted with water (4.5 L) and toluene
(6.0 L). The layers were separated, and the organic was
washed with water (2 × 3.0 L) and concentrated under
reduced pressure to give 1 as a yellow oil (541.5 g assay) in
97% yield as the free base. The crude oil was dissolved in
toluene (2.2 L), and anhydrous hydrogen chloride (4.5 N in
2-propanol, ca. 209 mL, 0.94 mol) was added over 5 min at
room temperature. n-Heptane (550 mL) was added in one
portion. The mixture was seeded and aged for 2 h at room
temperature. The remaining n-heptane (7.2 L) was added
slowly over 2 h, and the resulting slurry was further aged
for an additional 2 h at room temperature. The solids were
filtered, washed with a 3:1 mixture of n-heptane/toluene (1.5
L), and dried at 40 °C under vacuum with a nitrogen stream
to afford the pure hydrochloride salt of 1 (546.5 g) in 95%
yield as a white solid. The overall isolated yield for the entire
25
process was 89%. [R]D +163 (c 1.0, MeOH): mp 201-
1
202 °C (ethyl acetate/isopropyl alcohol/heptane). H NMR
(400 MHz, CD3CN) δ 7.75 (s, 1 H), 7.54 (br t, J ) 6.4 Hz,
2 H), 7.38 (s, 2 H), 7.11 (t, J ) 8.8 Hz, 2 H), 4.90 (q, J )
6.6 Hz, 1 H), 4.34 (d, J ) 2.8 Hz, 1 H), 4.21 (d, J ) 14.1
Hz, 1 H), 4.15 (d, J ) 14.1 Hz, 1 H), 4.10 (td, J ) 11.4 and
2.4 Hz, 1 H), 3.69 (d, J ) 13.9 Hz, 1 H), 3.56 (ddd, J )
11.4, 3.2, and 1.6 Hz, 1 H), 3.48 (br m, 1 H), 3.23 (d, J )
13.9 Hz, 1 H), 2.78 (s, 6 H), 2.74 (od, 1 H), 2.44 (br t, J )
12.0 Hz, 1 H), 1.44 (d, J ) 6.4 Hz, 3 H);13C NMR (100
MHz, CD3CN) δ 163.5 (d, J ) 244.0 Hz, Csp2-F), 147.2,
142.3, 135.4, 134.1, 132.7 (d, J ) 8.0 Hz), 131.9 (q, J )
32.9 Hz, Csp2-CF3), 127.9, 124.4 (q, J ) 271.5 Hz, CF3),
122.3, 116.0 (d, J ) 21.7 Hz), 96.3, 73.1, 70.2, 60.1, 58.2,
53.0, 50.7, 49.4, 42.9, 24.5. Anal. Calcd for C26H29ClF7N5O2
(611.98) C, 51.03; H, 4.78; N, 11.44. Found: C, 51.11; H,
4.67; N, 11.33.
4-N,N-Dimethylaminomethyl-5-formyl-1,2,3-triazole (3).
A solution of 1-N,N-(dimethylamino)-2-propyne (83.5 g, 1.0
mol) in THF (460 mL) was cooled to 0-5 °C, and
ethylmagnesium chloride (2.0 M, 550.0 mL, 1.1 mol) was
added over 20 min while maintaining the temperature
between 20 and 25 °C. The cold bath was removed, and the
Vol. 9, No. 4, 2005 / Organic Process Research & Development
•
497

Method C: Piperidine Dimer-DeriVatization Process. The
DMSO solution of triazole aldehyde (pH ≈ 9.7) was pH
adjusted to 7.5-8.5 with the dropwise addition of anhydrous
HCl in 2-propanol (4.2 N, ca. 460 mL, ca. 1.95 mol) over
15 min at room temperature. After completion of the addition,
the resulting yellow slurry was concentrated to remove THF
and 2-propanol. The mixture was filtered through a pad of
Celite and rinsed with DMSO (285 mL). Piperidine (108.5
mL, 1.1 mol) in ethyl acetate (985 mL, 1 M solution) was
added to the DMSO filtrate at room temperature over 2 h,
and the mixture was aged for a further 16 h. The resulting
slurry was filtered to afford product 25 as a white solid,
which was washed with DMSO (300 mL) and ethyl acetate
(500 mL), and dried at 40 °C under vacuum with a nitrogen
stream to afford the dimer adduct 25 (171.5 g assay) in 75%
yield based on 1-(dimethylamino)-2-propyne (79% relative
to sodium azide). The piperidine adduct hydrolyzed on the
analytical HPLC column and was assayed as triazole
aldehyde equivalent: mp 182-183 °C; 1H NMR (300 MHz,
CDCl3) δ 6.59 (s, 2 H), 3.97 (d, J ) 13.2 Hz, 2 H), 3.51 (d,
J ) 13.2 Hz, 2 H), 2.53-2.74 (m, 4 H), 2.35-2.47 (m, 4
H), 2.29 (s, 12 H); 1.32-1.59 (m, 12 H);13C NMR (75 MHz,
CDCl3) δ 142.4; 129.6; 73.6; 54.0; 48.8; 45.4; 26.0; 23.9.
Anal. Calcd for C22H40N10O (460.62) C, 57.37; H, 8.75; N,
30.41. Found: C, 59.32; H, 8.60; N, 31.50.
Piperidine adduct 25 (171.5 g assay, 0.373 mol) was
slurried in 2-propanol/water (98/2, 1.1 L), and trifluoroacetic
acid (57.4 mL, 0.745 mol) was added over 10 min at room
temperature with cooling. The triazole aldehyde was liberated
during the addition and crystallized as a white solid from
the reaction mixture. The slurry was stirred for 2 h at room
temperature. The solid was filtered, washed with 2-propanol
(150 mL), and dried at 40 °C under vacuum with a nitrogen
stream to afford pure 4-N,N-dimethylaminomethyl-5-formyl-
1,2,3-triazole (3) (96.5 g) in 84% yield (63% overall isolated
yield from 1-(dimethylamino)-2-propyne): mp 184-185
°C;1H NMR (400 MHz, D2O) δ 9.90 (s, 1 H), 4.42 (s, 2 H),
2.77 (s, 6 H);13C NMR (100 MHz, D2O) δ 188.3; 144.0;
137.6; 51.7; 42.6. Anal. Calcd for C6H10N4O (154.17) C,
46.74; H, 6.54; N, 36.34. Found: C, 46.50; H, 6.58; N, 36.61.
2-(R)-(1-(R)-(3,5-Bis(trifluoromethyl)phenyl)ethoxy)-4-
(5-(dimethylamino)methyl-1,2,3-triazol-4-yl)methyl-3-(S)-
(4-fluorophenyl)morpholine (1). Sodium triacetoxyboro-
hydride (95 wt %, 424.0 g, 1.9 mol) was added to anhydrous
N,N-dimethylacetamide (DMAC) (650 mL), and the mixture
was aged at room temperature for 0.5 h. The resulting cloudy
solution was added in one portion to a cold (0-5 °C) solution
of 2-(R)-(1-(R)-(3,5-bis(trifluoromethyl)phenyl)ethoxy)-3-(S)-
(4-fluorophenyl)morpholine p-toluenesulfonic salt (2) (609.0
g, 1.0 mol) and 4-N,N-dimethylaminomethyl-5-formyl-1,2,3-
triazole (3) (200.5 g, 1.3 mol) in DMAC (1.3 L). The mixture
was warmed to room temperature, heated at 40 °C for 2 h,
and cooled to room temperature. Aqueous hydrochloric acid
(2.0 N, ca. 1.75 L, ca. 3.5 mol) was added to bring the pH
down to 1.5-2.5. The mixture was heated again at 40 °C
for 2 h, cooled to room temperature, and diluted with toluene
(6.6 L). The aqueous layer was neutralized to pH ≈ 8.5-
9.0 with aqueous sodium hydroxide (5.0 N, ca. 1.9 L). The
layers were separated, and the organic phase was washed
with water (2 × 3.0 L) and concentrated under reduced
pressure to give 1 as a yellow oil (570.0 g assay) in 99%
yield as the free base. The material was converted to the
HCl salt without further purification. The crude oil was
dissolved in toluene (2.2 L), and anhydrous hydrogen
chloride (4.5 N in 2-propanol, ca. 215 mL, 0.97 mol) was
added over 5 min at room temperature. n-Heptane (550 mL)
was added in one portion. The mixture was seeded (0.5 wt
%, 3.0 g) and aged for 2 h with stirring at room temperature.
The remaining n-heptane (7.2 L) was added slowly over 2
h. The resulting slurry was aged for an additional 2 h at room
temperature with stirring. The solid was filtered, washed with
a 3:1 mixture of n-heptane/toluene (1.5 L), and dried at 40
°C under vacuum with a nitrogen stream to afford the pure
1 hydrochloride salt (575.5 g) in 95% yield as a white solid.
[R]D25 +163 (c 1.0, MeOH): mp 201-202 °C;1H NMR (400
MHz, CD3CN) δ 7.75 (s, 1 H), 7.54 (br t, J ) 6.4 Hz, 2 H),
7.38 (s, 2 H), 7.11 (t, J ) 8.8 Hz, 2 H), 4.90 (q, J ) 6.6 Hz,
1 H), 4.34 (d, J ) 2.8 Hz, 1 H), 4.21 (d, J ) 14.1 Hz, 1 H),
4.15 (d, J ) 14.1 Hz, 1 H), 4.10 (td, J ) 11.4 and 2.4 Hz,
1 H), 3.69 (d, J ) 13.9 Hz, 1 H), 3.56 (ddd, J ) 11.4, 3.2,
and 1.6 Hz, 1 H), 3.48 (br m, 1 H), 3.23 (d, J ) 13.9 Hz, 1
H), 2.78 (s, 6 H), 2.74 (od, 1 H), 2.44 (br t, J ) 12.0 Hz, 1
H), 1.44 (d, J ) 6.4 Hz, 3 H);13C NMR (100 MHz, CD3-
CN) δ 163.5 (d, J ) 244.0 Hz, Csp2-F), 147.2, 142.3, 135.4,
134.1, 132.7 (d, J ) 8.0 Hz), 131.9 (q, J ) 32.9 Hz, Csp2-
CF3), 127.9, 124.4 (q, J ) 271.5 Hz, CF3), 122.3, 116.0 (d,
J ) 21.7 Hz), 96.3, 73.1, 70.2, 60.1, 58.2, 53.0, 50.7, 49.4,
42.9, 24.5. Anal. Calcd for C26H29ClF7N5O2 (611.98) C,
51.03; H, 4.78; N, 11.44. Found: C, 51.11; H, 4.67; N, 11.33.
Received for review February 14, 2005.
OP050019S
498
•
Vol. 9, No. 4, 2005 / Organic Process Research & Development