A practical synthesis for the core structure of a family of selective prostaglandin D2 receptor antagonists

Article abstract of DOI:10.1021/jo026511g

A convergent synthesis was developed for the production of the core structure of prostaglandin D₂ receptor antagonists for the treatment of allergic rhinitis. The key steps in this synthesis were a highly diastereoselective alkylation of (+)-nopinone, a chemo- and stereoselective reduction of an oxime to an amine, and a well-controlled reduction of an aminoalkyne to a (Z)-olefin.

Full text of DOI:10.1021/jo026511g

A P r a ctica l Syn th esis for th e Cor e Str u ctu r e of a F a m ily of
Selective P r osta gla n d in D2 Recep tor An ta gon ists
Kevin R. Campos,* Michel J ournet,* Dongwei Cai,* J ason J . Kowal, Sandra Lee,
Robert D. Larsen, and Paul J . Reider
Department of Process Research, Merck Research Laboratories, P.O. Box, 2000, RY800-B267,
Rahway, New J ersey 07065-0900
kevin_campos@merck.com
Received October 2, 2002
A convergent synthesis was developed for the production of the core structure of prostaglandin D2
receptor antagonists for the treatment of allergic rhinitis. The key steps in this synthesis were a
highly diastereoselective alkylation of (+)-nopinone, a chemo- and stereoselective reduction of an
oxime to an amine, and a well-controlled reduction of an aminoalkyne to a (Z)-olefin.
In tr od u ction
protection/deprotection sequence of the amine through
2 and multiple operations to incorporate the side chain.
The primary goal in redesigning the synthesis was to
develop a streamlined, convergent route to 1 (Scheme 1).
Retrosynthetically, we envisioned that 1 could be derived
from ketoalkyne 3, which was the product of the alkyl-
ation of (+)-nopinone (4) with a propargyl halide such
as 5.5 This strategy was attractive because the synthesis
was highly convergent, and the starting materials were
either commercially available in bulk or readily prepared
on kilogram scale.
Prostaglandin D2 (PGD2) is the major cyclooxygenase
metabolite of arachidonic acid produced by mast cells in
response to antigen challenge.1 It has been proposed that
excess production of PGD2 causes the inflammation
commonly observed in allergic diseases such as allergic
rhinitis, asthma, and atopic dermatitis.2 Efforts to de-
velop a PGD2 receptor antagonist have identified the
class of compounds bearing the core structure 1 as
promising leads in the alleviation of various allergic
disorders.3 In this paper, the full details of the develop-
ment of a highly convergent, efficient synthesis of 1 are
disclosed.
Discu ssion
Alk yla tion of (+)-Nop in on e. Our initial efforts were
focused on the alkylation of the enolate of (+)-nopinone
(4) with a model propargyl bromide 5a . We quickly
discovered that the temperature of the reaction had a
dramatic effect on the yield and selectivity of the alkyl-
ation (Table 1). At higher temperatures, the yields were
respectable, but the diastereoselectivity was poor (72:28),
favoring the desired, kinetically preferred product 6a .6
Conversely, as the temperature was decreased, the
diastereomeric ratios were as high as 99:1 favoring 6a ,
but the yields dropped to as low as 15%.
These results indicated a highly diastereoselective,
kinetically controlled alkylation to afford 6a at low
temperature, which was competitive with epimerization
to the thermodynamically favored product 7a at higher
temperatures. To access 6a in good yield, the reactivity
The previously published synthesis of 1 utilized (+)-
myrtenol as the source of the pinane core structure and
involved several steps which were not amenable to a
scalable process.4 In addition, the synthesis required a
(1) Lewis, R, A.; Soter, N. A.; Diamond, P. T.; Austen, K. F.; Oates,
J . A.; Roberts, L. J . J . Immunol. 1982, 129, 1627-1631.
(2) (a) Matsuoka, T.; Hirata, M.; Tanaka, H.; Takahashi, Y.; Murata,
T.; Kabashima, K.; Sugimoto, Y.; Kobayashi, T.; Ushikubi, F.; Aze, Y.;
Eguchi, N.; Urade, Y.; Yoshida, N.; Kimura, K.; Mizoguchi, A.; Honda,
Y.; Nagai, H.; Narumiya, S.; Kato, M.; Watanabe, M.; Vogler, B.; Awen,
B.; Masuda, Y.; Tooyama, Y.; Yoshikoshi, A. Science 2000, 287, 2013-
2016. (b) Charlesworth, E. N.; Kagey-Sobotka, A.; Schliemer, R. P.;
Norman, P. S.; Lichtenstein, L. M. J . Immunol. 1991, 149, 671-676.
(c) Proud, D.; Sweet, J .; Stein, P.; Settipane, R. A.; Kagey-Sobotka, A.;
Freidlander, M.; Lichtenstein, L. M. J . Allergy Clin. Immunol. 1990,
85, 896-905. (d) Murray, J . J .; Tonnel, A. B.; Brash, A. R.; Roberts, L.
J .; Gosset, P.; Workman, R.; Capron, A.; Oates, J . N. Engl. J . Med.
1986, 315, 800-804.
(4) Seno, K.; Hagashita, S. Chem. Pharm. Bull. 1989, 37, 1524-
1533.
(5) Alkylation of ketones with this propargyl halide have been
extensively utilized in prostaglandin syntheses: (a) Chen, S.; J anda,
K. D. J . Am. Chem. Soc. 1997, 119, 8724-8725. (b) Sato, T.; Shima,
H.; Otera, J . J . Org. Chem. 1995, 60, 3936-3937. (c) Suzuki, M.;
Yanagisawa, A.; Noyori, R. J . Am. Chem. Soc. 1988, 110, 4718-4726.
(d) Corey, E. J .; Niimura, K.; Hashimoto, S.; Hamada, Y. Tetrahedron
Lett. 1986, 27, 2199-2202. (e) Binns, M. R.; Haynes, R. K.; Lambert,
D. E.; Schober, P. A. Tetrahedron Lett. 1985, 26, 3385-3388.
(6) Diastereomer 6a (stereochemistry determined by 1H NMR) was
treated under thermodynamic equilibration with NaOEt in EtOH to
yield 7a (stereochemistry confirmed by 1H NMR) as the major product.
The diastereomeric ratio of 6a :7a was determined to be 15:85 by HPLC.
(3) Tsuri, T.; Honma, T.; Hiramatsu, Y.; Mitsumori, S.; Inagaki, M.;
Arimura, A.; Yasui, K.; Asanuma, F.; Kishino, J .; Ohtani, M. J . Med.
Chem. 1997, 40, 3504-3507.
10.1021/jo026511g CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/19/2003
2338
J . Org. Chem. 2003, 68, 2338-2342

Synthesis of Prostaglandin D2 Receptor Antagonists
SCHEME 1
TABLE 1. Effect of Tem p er a tu r e on
Dia ster eoselectivity a n d Yield a
the model system (90% yield, 99:1 6c:7c). This alkylation
proved to be quite general, affording high yields and
excellent diastereoselectivities for a variety of carbon and
heteroatom electrophiles.10
Red u ctive Am in a tion . With alkylated adduct 6c in
hand, our efforts were focused on methodology suitable
to convert the ketone to a primary amine. Reductive
amination with ammonium acetate and NaBH3CN was
very sluggish and the chemoselectivity was solvent
dependent. For example, only reduction of the ketone to
the alcohol was observed in THF; however, in MeOH,
reductive amination occurred in very low conversion
(10%) and poor diastereoselectivity (eq 1). Other amines
were also investigated with similar difficulties in effecting
reductive amination. We concluded that the low reactivity
of the ketone could be attributed to the steric environ-
ment around the reacting center.11
temp (°C)
conversion (%)b
6a :7a c
0
-25
-35
-45
-55
85
68
50
30
15
72:28
90:10
96:4
99:1
99:1
a
All reactions were carried out in THF (1.2 M in substrate),
b
using 1.05 equiv of LDA and 1.05 equiv of 5a . Conversion was
determined 1.5 h after addition of 5a (HPLC). c Diastereoselec-
tivity was determined by HPLC.
TABLE 2. Dia ster eoselective Nop in on e Alk yla tion w ith
Ca r bon Electr op h ilesa
Although the one-step reductive amination procedure
was unsuccessful, the two-step approach (oxime forma-
tion, reduction) was considered a reasonable alternative.
The methyl oxime 8 was formed in good yield (88%, eq
2); however, conditions for the reduction of 8 to the amine
were not anemable to the functionality present in the side
chain. Hydrogenation conditions selectively reduced the
alkyne before the oxime, and hydride reagents (LiAlH4,
DiBAL-H) reduced the ester first. Milder conditions (Zn/
TFA,12 MoO3/NaBH413) afforded no reaction.
a
All reactions were carried out in THF (1.2 M in substrate),
b
using 1.05 equiv of LDA and 1.05 equiv of RX. Assay yield of
the reaction was determined by HPLC. c Diastereoselectivity was
1
determined by HPLC and stereochemistry was determined by H
NMR.
of the enolate or the electrophile needed to be increased
such that the alkylation would occur at -45 °C. When
the alkylation with 5a was performed with 1 equiv of
DMPU at -45 °C, a high yield and high diastereoselec-
tivity were observed (90% yield, 99:1, 6a :7a ).7
Alternatively, when the more reactive propargyl iodide
5b (Table 2) was used in the alkylation in the absence of
DMPU, the same high yield and diastereoselectivity was
obtained (90% yield, 99:1 6a :7a ). In every case, the
reaction was quenched at low temperature (-45 °C) with
trifluoroacetic acid (TFA) to avoid epimerization to the
thermodynamically favored product.8 These optimized
conditions were then applied to the alkylation of (+)-
nopinone with propargyl iodide 5c,9 affording alkylated
adduct 6c with the same results that were observed in
Whereas 8 was chemically inert to MoO3/NaBH4, oxime
9 was reduced to the amine under these conditions in
(8) When the reaction was quenched by direct addition of aqueous
acid, some epimerization to the thermodynamically favored product
was always observed. In contrast, the reaction can be quenched by
direct addition of trifluoroacetic acid (TFA) to the reaction solution at
-45 °C or by reverse quench of the reaction solution at -45 °C into a
1:1 mixture of MTBE:satd NH4Cl(aq) with no epimerization.
(9) Prepared according to the method of Corey and Sachdev: Corey,
E. J .; Sachdev, H. S. J . Am. Chem. Soc. 1973, 95, 8483-8484.
(10) Campos, K.; Lee, S.; J ournet, M.; Kowal, J .; Cai, D.; Larsen,
R.; Reider, P. Tetrahedron Lett. 2002, 43, 6957-6959.
(7) Other additives (HMPA, TMEDA) were also investigated but
none were as effective as DMPU. Alkylation is complete within 1.5 h
at -45 °C. Lower temperatures afforded sluggish reactions.
J . Org. Chem, Vol. 68, No. 6, 2003 2339

Campos et al.
less than 1 h (eq 3). Additionally, partial reduction of the
alkyne was also observed (25%, 3:1:1 cis:trans:alkane).
The lability of 9 provided the necessary reactivity that
was required to successfully reduce the oxime in a
chemoselective fashion.
the rejection of 12 was very difficult in the purification
of the final drug, we sought a process for the well-
controlled hydrogenation of aminoalkyne 10.
Reduction of the alkene intermediates could be over-
come via protection of the primary amine moiety with a
carbonyl-containing protecting group. The Boc-protected
amine 13 could be reduced with Lindlar’s catalyst in
acetone by using low catalyst loading (6 wt %) to produce
the desired N-protected aminoalkene 2a with only 0.5%
of the alkane impurity. This result suggested that the
unprotected amine functionality was playing a role in the
overreduction pathway.18
It is well-documented that TiCl3 is an excellent reagent
for the mild reduction of N-O bonds.14 This reagent has
been used in conjuction with a mild reducing agent to
effect the reduction of oximes to amines in a one-pot
procedure.14,15 Treatment of substrate 9 under typical
conditions (NaBH3CN, TiCl3, MeOH) afforded complete
conversion to the aminoalkyne product 10 with good
diastereoselectivity (98:2) and, as expected, no reduction
of the alkyne. Due to the safety concerns and costliness
of handling a cyanide-containing waste stream on large
scale, we sought an alternative reducing agent to NaBH3-
CN.16 When NaBH(OAc)3 was substituted for NaBH3CN
as the reducing agent, only reduction of the N-O bond
was observed, delivering the imine; however, using
t-BuNH2-BH3 instead of NaBH3CN cleanly produced
aminoalkyne 10, which was then isolated by crystalliza-
tion as the HCl salt in 85% overall yield.
Closer investigation of the hydrogenation of 10 at lower
catalyst loadings (4 wt %) revealed that overreduction
was only significant after the alkyne was completely
consumed.19 Although it is theoretically possible to stop
the hydrogenation at a point where 10 is consumed and
12 has not been formed in appreciable amounts, this
solution is irreproducible, unreliable, and poorly con-
trolled. A more reliable solution to the problem was to
attempt to control overreduction through the use of an
additive.
The fact that the rate of reduction of 1 only becomes
significant after complete consumption of 10 suggests
that the preassociation of the catalyst surface is stronger
to 10 than to 1.20 One might expect, based on this model,
that the most suitable additive to attenuate the reactivity
of 1 would be a bidentate ligand to compete for catalyst
surface association. As expected, monodentate amines
such as pyridine or quinoline were ineffective at prevent-
ing overreduction (Table 3);21 however, a number of
bidentate amines were highly effective, reducing the
amount of 12 to e0.7%.22 Of the diamines investigated,
(18) Significant overreduction has been observed with other sub-
strates containing unprotected amines. (a) Lindel, T.; Hochguertel, M.
Tetrahedron Lett. 1998, 39, 2541-2544. (b) Sagi, G.; Oetvoes, L.; Ikeda,
S.; Andrei, G.; Snoeck, R.; De Clerq, E. J . Med. Chem. 1994, 37, 1307-
1311. (c) Merino, P.; Castillo, E.; Franco, S.; Merchan, F. L.; Tejero, T.
Tetrahedron: Asymmetry 1998, 9, 1759-1769. (d) Cliff, M. D.; Pyne,
S. G. J . Org. Chem. 1997, 62, 1023-1032.
(19) With only 4 wt % catalyst, the hydrogenation was complete in
less than 2 h. At the end of the reaction, the ratio of 4b:4c:4d was
94:4:2; however, by the time that the ratio had been determined by
HPLC (25 min), the ratio had already changed to 91:5:4.
Hyd r ogen a tion . On the basis of literature precedent,
the partial hydrogenation of the aminoalkyne was per-
formed with Lindlar’s catalyst.17 The starting material
was completely consumed in less than 1 h with low
catalyst loadings (10 wt %); however, analysis of the
reaction mixture showed a significant amount of both the
trans isomer 11 (8.5%) and the fully reduced product 12
(14%). Lower catalyst loadings (6 wt %) did slow the rate
of hydrogenation; however, the amount of alkane was still
at an unacceptably high level (6%) by the time that the
starting material had been completely consumed. Since
(20) The preassociation of an amine to a catalyst surface (Pd/C)
has been previously reported to account for surprising selectivities in
olefin hydrogenations. See: Thompson, H. W.; Wong, J . K. J . Org.
Chem. 1985, 50, 4270-4276. This is in accord with the relative
coordination strength of an alkyne to palladium versus that of an olefin.
Typically, the metal-carbon bonds in alkyne complexes are 0.1 Å
shorter than those with their olefin counterparts. (a) Redhouse, A. D.
In The Chemistry of the Metal-Carbon Bond; Hartley, F. R., Patai,
S., Eds.; J ohn Wiley and Sons: New York, 1982; Vol. 1, Chapter 1, pp
18-20. (b) Davies, J . A. In Comprehensive Organometallic Chemistry;
Puddephatt, R. J ., Ed.; Pergamon Press: New York, 1986; Vol. 9,
Chapter 6, pp 293-320.
(11) For an excellent review of reductive amination of a variety of
ketones see: Abdel-Magid, A.; Carson, K.; Harris, B.; Maryanoff, C.
A.; Shah. R. J . Org. Chem. 1996, 61, 3849-3862.
(12) Scolastico, C.; Conca, E.; Prati, L.; Guanti, G.; Banfi, L.; Berti,
A.; Farina, P.; Valcavi, U. Synthesis 1985, 850-855.
(13) Ipaktschi, J . Chem. Ber. 1984, 117, 856-858.
(21) Diphenylphosphinoethane was used as a catalyst poison as well
(0.1 mol %); however, this was too effective a poison, affording no
hydrogenation over 24 h.
(14) Hoffman, C.; Tanke, R. S.; Miller, M. J . J . Org. Chem. 1989,
54, 3750-3751 and references cited therein.
(15) Leeds, J . P.; Kirst, H. A. Synth. Commun. 1988, 18, 777-782.
(16) Additional safety concerns were raised regarding the potential
hazard of using NaBH3CN under acidic conditions on large scale (TiCl3
comes as a 20% solution in aqueous HCl).
(17) Lindlar, H.; Dubuis, R. Organic Syntheses; Wiley: New York,
1973; Collect. Vol. V, pp 880-883.
(22) The hydrogenations were complete in 2-3 h; however, to test
the robustness of the process, the reactions were allowed to stir for 18
h with no increase in the amount of the aminoalkane impurity. The
catalyst level could be increased up to 10 wt % with very little change
in the amount of overreduction observed (4 wt % 0.7% alkane; 10 wt
% 0.9% alkane).
2340 J . Org. Chem., Vol. 68, No. 6, 2003

Synthesis of Prostaglandin D2 Receptor Antagonists
TABLE 3. Sem ih yd r ogen a tion of Am in oa lk yn e 10 w ith
Am in e Ad d itivesa
6.34 L, 15.2 mol) was added over 45 min while maintaining
the temperature below -10 °C. The solution was aged for 10
min and (1R)-(+)-nopinone (2.00 kg, 92% optically pure, 14.5
mol, neat) was added over 45 min while keeping the temper-
ature below -10 °C. After complete addition, the solution was
aged for 30 min and cooled to -50 °C. The iodo-ester side-
chain 5c (4.25 kg assay, 15.2 mol) was added over 1.5 h while
keeping the temperature between -48 and -45 °C. Upon
complete addition, the reaction mixture was aged for 1.5 h at
-45 °C. Trifluoroacetic acid (1.68 L, 21.7 mol) was then added
over 1 h maintaining the temperature below -45 °C. The
reaction mixture was then allowed to warm to -10 °C and
transferred into a mixture of ethyl acetate (20.0 L) and 3%
aqueous L-tartaric acid (20.0 L). The layers were separated
and the organic was washed with water (2 × 20.0 L). The
organic layer was concentrated to an orange oil that was used
directly in the next step. Assay yield ) 3.90 kg (92%). A sample
entry
additive
pyridine
1:11:12c
1
2
3
4
5
6
70:20:10
80:13:7
97.6:2.0:0.4
96.7:2.5:0.8
97.0:2.5:0.5
97.3:2.3:0.4
quinoline
ethylenediamine
propylenediamine
1,10-phenanthroline
diethylenetriamine
a
All reactions were carried out in DMF (0.5 M in substrate) at
b
rt. Ratio of products determined by HPLC analysis.
of 6c was purified by flash chromatography (90:10 hexane:
23
ethyl acetate) to afford an analytically pure sample. [RD
]
ethylenediamine (EDA) afforded the least overreduction
and Z/E isomerization, providing the desired cis amino-
alkene 1 in >97% yield with only 2% isomerization to
the trans isomer 11 and 0.4% overreduction to 12.23 This
methodology was demonstrated to be effective for a
variety of aminoalkynes of varying tether length between
the amine and alkyne functionalities.24
For ease of handling, 1 was readily crystallized from
pure heptane as the HCl salt to afford material that is
>98 area % pure by HPLC. We were delighted to find
that the enantiomeric excess of the crystallized material
had been upgraded from 92% (technical grade â-pinene)
to 99%. This intermediate can be readily coupled with a
variety of sulfonylating and acylating agents to form the
family of PGD2 receptor antagonists 2 in good yield and
high enantiomeric excess.
+62.7 (c 0.10, 92% ee, CCl4); IR (film) 2976, 2938, 2872, 1734,
1710, 1456-997 cm-1; 1H NMR (400 MHz, CDCl3) 4.12 (q, 2H,
J ) 7.2 Hz), 2.65-2.49 (m, 5H), 2.39 (t, 2H, J ) 7.8 Hz), 2.28-
2.17 (m, 4H), 1.98-1.93 (m, 1H), 1.78 (quint, 2H, J ) 7.3 Hz),
1.58 (d, 1H, J ) 9.6 Hz), 1.33 (s, 3H), 1.25 (s, 3H), 0.91 (s,
1H); 13C NMR (400 MHz, CDCl3) δ 14.1, 18.1, 22.0, 24.0, 24.3,
25.7, 28.1, 33.0, 40.3, 40.9, 42.4, 58.0, 60.2, 78.6, 80.6, 173.1,
213.6; TLC Rf 0.35 (80% hexane, 20% EtOAc); separation of
diastereomers was accomplished by HPLC analysis (Waters
Symmetry C18, 0.1% HClO4:MeCN, 20:80) to show a diastereo-
meric ratio of 99:1 (6c:7c). Anal. Calcd for C18H26O3: C, 74.45;
H, 9.02. Found: C, 74.27; H, 8.62.
Eth yl 7-[(1R,2E,Z,3R,5S)-2-(Hyd r oxyim in o)-6,6-d im eth -
ylbicyclo[3.1.1]h ep t-3-yl]h ep t-5-yn oa te (9). A 50-L round-
bottom flask was charged with 6c (3.90 kg assay, 13.4 mol),
ethanol (13.0 L), and water (6.4 L). Sodium acetate trihydrate
(3.63 kg, 26.7 mol) was added followed by hydroxylamine
hydrochloride (1.86 kg, 26.7 mol). The reaction mixture was
then heated to 50 °C (homogeneous) for 5 h. (Note: It is well-
known that hydroxylamine hydrochloride should be stored
away from any heat source due to the significant differential
scanning calorimetry (DSC) exotherm at ∼100 °C. Operational
Hazards Evaluation of the oxime formation showed that a
reaction temperature of 50 °C was suitable for the preparation
of the oxime on multikilogram scale.) The reaction was cooled
to room temperature and pumped into toluene (20.0 L) and
water (20.0 L). The layers were separated and the organic layer
was washed with water (2 × 20.0 L). The organic layer was
then concentrated to an orange oil comprised of a mixture of
E,Z isomers of the oxime that was used directly in the next
step. Assay yield 3.67 kg (90%). IR (film) 3241(br), 2933(br),
2871, 1734, 1653, 1457, 1371, 1159, 1028, 948 cm-1; 1H NMR
(400 MHz, CDCl3) 4.14 (q, 2H, J ) 7.1 Hz), 2.59-2.55 (m, 1H),
2.49-2.38 (m, 5H), 2.28-2.20 (m, 4H), 2.15-2.06 (m, 2H),
2.01-1.94 (m, 1H), 1.89-1.77 (m, 2H), 1.51 (d, 1H, J ) 10.4
Hz, major isomer), 1.39 (d, 1H, J ) 1.6 Hz, minor isomer), 1.34
(s, 3H, minor isomer), 1.28 (s, 3H, major isomer), 1.26 (t, 6H,
J ) 7.0 Hz), 0.85 (s, 3H, minor isomer), 0.77 (s, 3H, major
isomer); 13C NMR (400 MHz, CDCl3) δ 14.1, 18.2, 21.5, 24.1,
24.9, 25.4, 25.6, 28.6, 30.8, 33.1, 40.8, 41.6, 48.2, 60.2, 79.4,
80.0, 165.4, 173.2; TLC (85% petroleum ether, 15% EtOAc);
Rf(major) 0.27; Rf(minor) 0.20; separation of E and Z isomers was
accomplished by HPLC analysis (Waters Symmetry C18, 0.1%
HClO4:MeCN, 20:80) to show a 58:42 mixture of isomers. Anal.
Calcd for C18H27NO3: C, 70.79; H, 8.91; N, 4.59. Found: C,
70.63; H, 8.97; N, 4.71.
Su m m a r y
In conclusion, a highly convergent synthesis of 1 has
been described that has significant advantages over
previous routes with respect to ease of scale-up and
convergence, allowing for the production of a variety
amide derivatives from the same common intermediate.
These studies have also resulted in the development of
methodology for a highly diastereoselective alkylation of
(+)-nopinone, a mild, chemo- and stereoselective reduc-
tion of an oxime, and a well-controlled semihydrogenation
of aminoalkynes.
Exp er im en ta l Section
NMR spectra (1H, 13C) were recorded with CDCl3 as solvent.
The reactions and products were assayed by HPLC with water
and acetonitrile as eluting solvents. Diisopropylamine, ethyl-
enediamine, and all solvents were used without further
purification. Titanium trichloride was used as a 20% solution
in 3% aqueous HCl. Hydrogenations were run in a pressurized
vessel, degassed by vacuum, and then purged with hydrogen.
Eth yl 7-[(1R,3R,5S)-6,6-Dim eth yl-2-oxobicyclo[3.1.1]-
h ep t-3-yl]h ep t-5-yn oa te (6c). A 50-L round-bottom flask was
charged with dry THF (12 L) and diisopropylamine (2.23 L,
15.9 mol) and cooled to -30 °C. n-Butyllithium (2.4 M/Hex,
Eth yl 7-[(1R,2R,3R,5S)-2-Am in o-6,6-d im eth ylbicyclo-
[3.1.1]h ep t-3-yl]h ep t-5-yn oa te HCl Sa lt (10-HCl). To a
solution of aqueous TiCl3 (32 L, 20% in 3% HCl, 44.8 mol) at
25 °C was added sodium acetate (17.4 kg, 44.8 mol) with
stirring until complete dissolution was achieved. The solution
was cooled to 0 °C, and oxime 9 (3.90 kg, 12.8 mol) was added
as a solution in 32 L of ethanol while maintaining a temper-
(23) Ethylenediamine has been reported to be a catalyst poison for
other Pd catalysts. (a) Hattori, K.; Sajiki, H.; Hirota, K. Tetrahedron
Lett. 2000, 41, 5711-5714. (b) Enders, D.; Schankat, J . Helv. Chim.
Acta 1995, 78, 970-992. (c) Savoia, D.; Trombini, C.; Umani-Ronchi,
A.; Verardo, G. J . Chem. Soc., Chem. Commun. 1981, 540-541.
(24) Campos, K. R.; Cai, D.; J ournet, M.; Kowal, J . J .; Larsen, R.
D.; Reider, P. J . J . Org. Chem. 2001, 66, 3634-3635.
J . Org. Chem, Vol. 68, No. 6, 2003 2341

Campos et al.
ature of 0 °C. After the mixture was stirred for 1 h at 0 °C,
borane tert-butylamine complex (2.23 kg. 25.6 mol) was
charged over 10 min, keeping the temperature below 10 °C.
The resulting solution was aged for 30 min, and then diluted
with 64 L of ethyl acetate. The layers were separated, and the
organic layer was washed with saturated ammonium chloride
(32 L). The organic layer was neutralized with 1 M Na2CO3
until the pH of the aqueous solution was around 8. The organic
layer was washed once more with water and then concentrat-
ed,the solvent was switched to isopropyl acetate (13 L), and
the mixture was azeotropically dried. The resulting solution
of free base in isopropyl acetate was treated with HCl in
diethyl ether (2.0 M, 7.70 L) and aged for 30 min. The reaction
mixture was concentrated to remove ether and any excess HCl,
and the reaction volume was adjusted to 20 L with additional
isopropyl acetate. The reaction mixture was seeded with the
desired product followed by an addition of n-heptane (20 L)
over 1 h. After thge solution was stirred for 1 h, the slurry
was filtered, and the cake was washed with a 1:1 mixture of
isopropyl acetate and n-heptane (6 L). The cake was dried to
afford 10-HCl as a crystalline solid (4.59 kg, 3.90 kg assay,
85% isolated yield). Mp 137.5-138.5 °C; [RD23] +44.7 (c 0.01,
CCl4); IR (film) 2906 (br), 2877(br), 1734, 1602, 1505, 1376,
for 1 h, then the phases were separated. The organic layer
was washed with Na2CO3 (1 M, 6.9 L), then with H2O (14 L).
The organic layer was concentrated to yield 10 as an oil (3.43
kg assay by HPLC, 98%). The crude material was diluted with
DMF (9.4 L) followed by ethylenediamine (0.480 L, 7.19 mol),
and then Lindlar’s catalyst (5 wt % Pd on CaCO3 poisoned with
lead, 65 g). The reaction was subjected to hydrogen atmosphere
(40 psi) for 6 h. The mixture was filtered over solka floc, and
the filter bed was washed with 70:30 n-heptane:ethyl acetate
(5 L). The filtrate was diluted with 70:30 n-heptane:ethyl
acetate (20 L), washed once with 1.7% NH4Cl(aq) (35 L), and
washed twice with water (26 L). The organic solution was
distilled at constant volume (∼20 L) with pure n-heptane until
the Kf of the solution was <500 µg/mL. The reaction mixture
was cooled to 0 °C, and HCl in Et2O (2.2 M, 6.4 L) was added.
Upon complete addition, the reaction was allowed to warm to
RT and aged for 10 min. The mixture was distilled at constant
volume (∼35 L) with pure n-heptane to remove excess diethyl
ether and HCl. During the distillation, the HCl salt begins to
precipitate. The mixture was heated to 60 °C or until complete
dissolution of solid, and then allowed to cool to rt. After being
stirred for 4 h, the resulting slurry was filtered to yield 1-HCl
as a crystalline solid (3.52 g, 3.46 kg assay, 94% isolated yield).
Mp 90.9-91.8 °C; [RD23] +31.5 (c 0.012, CCl4); IR (film)
2994-2869, 1730, 1598, 1509 cm-1; 1H NMR (400 MHz, CDCl3)
5.53-5.43 (m, 2H), 4.12 (q, 2H, J ) 7.2 Hz), 3.37 (m, 1H),
2.54-2.48 (m, 1H), 2.41-2.38 (m, 2H), 2.31 (t, 2H, J ) 7.8
Hz), 2.25-2.12 (m, 4H), 1.70 (quint, 2H, J ) 7.4 Hz), 1.50 (m,
1H), 1.26 (s, 3H), 1.25 (t, 3H, J ) 7.2 Hz), 1.22 (s, 3H), 0.84
(br d, 1H, J ) 9.2 Hz); 13C NMR (400 MHz, CDCl3) 14.2, 23.0,
24.8, 26.7, 26.8, 30.6, 32.2, 33.7, 34.2, 36.4, 37.4, 40.8, 44.9,
58.6, 60.1, 127.2, 131.5, 172.5; separation of diastereomers was
accomplished by GC analysis HP-35 (cross-linked 35% PhME
Siloxane) to show a purity of 98.8% (0.5% 11, 0.2% 12). Anal.
Calcd for C18H32ClNO2: C, 65.53; H, 9.78; N, 4.25. Found: C,
65.57; H, 9.64; N, 4.20.
1
1151 cm-1; H NMR (400 MHz, CDCl3) 4.14 (q, 2H, J ) 7.1
Hz), 3.60-3.53 (m, 1H), 2.71-2.62 (m, 1H), 2.51-2.35 (m, 5H),
2.28-2.21 (m, 4H), 1.99-1.96 (m, 1H), 1.81 (quint, 2H, J )
7.2 Hz), 1.50 (m, 1H), 1.29 (s, 3H), 1.27 (t, 3H, J ) 7.2 Hz),
1.23 (s, 3H), 1.09 (br d, 1H, J ) 9.6 Hz); 13C NMR (400 MHz,
CDCl3) δ 14.1, 18.2, 22.9, 24.1, 25.9, 26.8, 30.0, 31.4, 33.2, 37.5,
40.6, 44.6, 57.4, 60.2, 76.6, 78.1, 81.2, 173.1; separation of
diastereomers was accomplished by HPLC analysis (Supelcosil
ABZ Plus, 0.1% HClO4:MeCN, 20:80) to show a diastereomeric
excess of 94%. Anal. Calcd for C18H30ClNO2: C, 65.93; H, 9.22;
N, 4.27. Found: C, 65.94; H, 9.14; N, 4.26.
E t h yl (5Z)-7-[(1R,2R,3R,5S)-2-Am in o-6,6-d im et h ylb i-
cyclo[3.1.1]h ep t-3-yl]h ep t-5-en oa te HCl Sa lt (1-HCl). To
an extractor containing isopropyl acetate (18 L) and the
aminoalkyne HCl salt 10-HCl (4.59 kg, 3.90 kg assay, 11.9
mol) was added Na2CO3 (1 M, 14 L). The mixture was stirred
J O026511G
2342 J . Org. Chem., Vol. 68, No. 6, 2003