Efficient asymmetric synthesis of N-[(1R)-6-chloro-2,3,4,9-tetrahydro-1H- carbazol-1-yl]-2-pyridinecarboxamide for treatment of human papillomavirus infections

Article abstract of DOI:10.1021/op060223v

An efficient asymmetric synthesis of N-(1R)-6-chloro-2,3,4,9-tetrahydro-1H- carbazol-1-yl]-2-pyridinecarboxamide 1, a potential treatment for human papillomavirus infections, is described. The key step in the synthesis of this molecule is an asymmetric reductive animation directed by chiral (phenyl)-ethylamines resulting in up to 96% disastereo facial selectivity. The synthesis is also highlighted by isolation of a unique 2-picolinic acid salt of (1R)-6-chloro-2,3,4,9-tetrahydro-1H-carbazol-1-amine (13). Subsequent application of 1-propylphosphonic acid cyclic anhydride (T3P) for convenient amide formation from the two components of the salt provides the product 1 in high yield. The process research work leading to the final synthesis includes a racemic synthesis followed by resolution with chiral supercritical fluid chromatography, and an enantioselective reductive animation via chiral transfer hydrogenation catalyzed by Ru(II) complexes of N-[(1S,2S)-2-amino-1,2- diphenylethyl]-1-naphthalenesulfonamide or (R)-BI-NAP. Highlighting the practicality of the synthesis, the process has been scaled up in 200-gallon reactors for delivery of multikilograms of the target compound 1 in over 99.5% enantiomeric purity.

Full text of DOI:10.1021/op060223v

Organic Process Research & Development 2007, 11, 539−545
Efficient Asymmetric Synthesis of
N-[(1R)-6-Chloro-2,3,4,9-tetrahydro-1H-carbazol-1-yl]-2-pyridinecarboxamide for
Treatment of Human Papillomavirus Infections
†
‡
,†
‡
‡
‡
Sharon D. Boggs, Jeremy D. Cobb, Kristjan S. Gudmundsson,* Lynda A. Jones, Richard T. Matsuoka, Alan Millar,
‡
†
‡
,‡
‡
Daniel E. Patterson, Vicente Samano, Mark D. Trone, Shiping Xie,* and Xiao-ming Zhou
Chemical DeVelopment and Medicinal Chemistry, GlaxoSmithKline,
Research Triangle Park, North Carolina 27709, U.S.A.
Abstract:
Papillomaviruses that cause genital infections are classi-
fied as low-risk or high-risk HPVs. The low-risk HPVs (such
3
An efficient asymmetric synthesis of N-[(1R)-6-chloro-2,3,4,9-
tetrahydro-1H-carbazol-1-yl]-2-pyridinecarboxamide 1, a po-
tential treatment for human papillomavirus infections, is
described. The key step in the synthesis of this molecule is an
asymmetric reductive amination directed by chiral (phenyl)-
ethylamines resulting in up to 96% disastereo facial selectivity.
The synthesis is also highlighted by isolation of a unique
as HPVs 6 and 11) cause genital warts, while the high-risk
HPVs (such as HPVs 16 and 18) are highly associated with
the development of genital cancers including cervical
carcinoma. The role of HPV as the principle agent in the
4
etiology of cervical cancer has been clearly established with
a lifetime risk of invasive cancer in the range of 5-10% for
untreated infections. While most HPV infections are tran-
sient, lasting less than 1 year, a high proportion of persistent
infections with high-risk types may progress to cervical
dysplasia, the precursor to cervical cancer.
2-picolinic acid salt of (1R)-6-chloro-2,3,4,9-tetrahydro-1H-
carbazol-1-amine (13). Subsequent application of 1-propylphos-
phonic acid cyclic anhydride (T3P) for convenient amide
formation from the two components of the salt provides the
product 1 in high yield. The process research work leading to
the final synthesis includes a racemic synthesis followed by
resolution with chiral supercritical fluid chromatography, and
an enantioselective reductive amination via chiral transfer
hydrogenation catalyzed by Ru(II) complexes of N-[(1S,2S)-2-
amino-1,2-diphenylethyl]-1-naphthalenesulfonamide or (R)-BI-
NAP. Highlighting the practicality of the synthesis, the process
has been scaled up in 200-gallon reactors for delivery of
multikilograms of the target compound 1 in over 99.5%
enantiomeric purity.
Currently available treatment for genital warts and cervical
dysplasia involve surgical removal or chemical destruction
of the infected tissue. The immunomodulator imiquimod
(Aldara) has been approved for the treatment of external
genital warts, but present treatments do not target the viral
infection. Recently, vaccines (Cervarix and Gardasil) have
been developed to prevent HPV infection. How effective
these will be in reducing new HPV infections remains to be
5
seen.
6
A novel series of tetrahydrocarbazoles was identified via
7
a cell-based screen. Optimization of these tetrahydrocarba-
zoles resulted in identification of 1 as a suitable development
candidate. A large supply of 1 was required to support the
preclinical and clinical development work. Herein we report
our process research work on three generations of synthetic
routes to 1: a racemic synthesis followed by a resolution
through chiral chromatography, an enantioselective synthesis
through chiral catalysis, and a chiral auxiliary directed
Introduction
Human papillomaviruses (HPVs) are small, non-envel-
oped DNA viruses that cause a wide variety of benign and
premalignant epithelial tumors. Several HPVs infect the
genital mucosa. These are sexually transmitted with an
incidence roughly twice that of Herpes simplex infection.
HPV infection is considered the most common sexually
(
3) (a) Syrjanen, K.; Mantyjarvi, R.; Saarikoski, S.; Vayrynen, M.; Syrjanen,
S.; Parkkinen, S.; Yliskoski, M.; Saastamoinen, J.; Castren, O. Br. J. Obstet.
Gynaecol. 1988, 95, 1096. (b) Wilson, V. G.; Rosas-Acosta, G. Curr. Drug
Targets: Infect. Disord. 2003, 3, 221. (c) Severson, J.; Evans, T. Y.; Lee,
P.; Chan, T.-S.; Arany, I.; Tyring, S. K. J. Cutaneous Med. Surg. 2001, 5,
43.
1
transmitted disease throughout the world. There are over
5.5 million new cases of sexually transmitted HPV in the
United States each year, with at least 20 million people
currently infected.2
(4) (a) Bosch, F. X.; Lorincz, A.; Munoz, N.; Meijer, C. J. L. M.; Shah, K. V.
J. Clin. Pathol. 2002, 55, 244. (b) zur Hausen, H. Nat. ReV. 2002, 2, 342.
*
To
whom
correspondence
should
be
addressed.
E-mail:
(5) (a) Snoeck, R. AntiViral Res. 2006, 71, 181. (b) Schmiedeskamp, M. R.;
Kockler, D. R. Ann. Pharmacother. 2006, 40, 1344.
kristjan.s.gudmundsson@gsk.com (K.S.G.); shiping.x.xie@gsk.com (S.X.).
†
Medicinal Chemistry.
Chemical Development.
(6) For previous literature on the synthesis of the tetrahydrocarbazole core
see: (a) Chakraborty, D. P.; Islam, A.; Bhattacharyya, P. J. Org. Chem.
1973, 38, 2728. (b) Chakraborty, D. P.; Chowdhury, B. K. J. Org. Chem.
1968, 32, 1265. (c) Shoeb, A.; Anwer, F.; Kapil, R. S.; Popli, S. P. J. Med.
Chem. 1973, 16, 425. (d) Mahboobi, S.; Kuhr, S.; Meindl, W. Arch. Pharm.
1994, 327, 611. (e) Balamurali, R. Rajendra Prasad, K. J. Il Farmaco 2001,
56, 229.(f) Borsche, W.; Witte, A.; Bothe, W. Justus Liebigs Ann. Chem.
1908, 359, 49. (g) Borsche, W.; Kienitz, G. A. Chem. Ber. 1910, 43, 2335.
(7) Boggs, S. D.; Gudmundsson, K. WO 05/005386, 2005.
‡
(
1) (a) Howley, P. M. In Fundamental Virology; Howley, P. M., Ed.; Lippincott-
Raven Publishers: Philadelphia, PA, 1996; pp 947-978. (b) Kroutsky, L.
Am. J. Med. 1997, 102, 3.
(
2) (a) Wilson, F. X. Expert Opin. Emerging Drugs 2001, 6, 199. (b) Department
of Health and Human Services Center for Disease Control and Prevention.
Genital HPV infection: CDC factsheet. www.cdc.gov/std/HPV/STDFact-
HPV.htm#common (accessed June 2006).
1
0.1021/op060223v CCC: $37.00 © 2007 American Chemical Society
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
539
Published on Web 02/14/2007

diastereoselective synthesis. The process research led to a
robust and efficient synthesis of 1 on multikilogram scale.
solubility of racemic 2 in the mobile phase hampered our
efforts to scale up the chiral SFC separation. While separation
of racemic 1 was also possible by SFC chromatography,12
the solubility of racemic 1 in the mobile phase was even
worse than that of the racemic amine precursor 2. We also
prepared some other N-derivatives such as triisopropylsilyl
and trifluoroacetyl analogues of racemic 2, and attempted
the chiral separation utilizing simulated moving bed (SMB)
chromatography on a chiral stationary phase. Unfortunately,
separation of these N-derivatives proved unsatisfactory due
to poor separation or low solubility.
Results and Discussion
The target molecule (1) is a 2-picolinamide of (R)-6-
chloro-2,3,4,9-tetrahydro-1H-carbazol-1-amine. While a num-
ber of syntheses of substituted 2,3,4,9-tetrahydro-1H-
Scheme 1
6
carbazole exist, our efforts focused on obtaining the chiral
amine 2 via a reductive amination of the ketone 3 (Figure
1). Picolinic acid or any other carboxylic acid could then be
incorporated in the last synthetic step. This strategy was
aligned with the structure-activity studies in the drug
discovery program which explored various different tetrahy-
drocarbazole amides. 6-Chloro-2,3,4,9-tetrahydro-1H-carba-
zol-1-one (3) could be prepared from 4-chloroaniline (4) and
2
2
-(hydroxymethylene)cyclohexanone (5),8 a tautomer of
-formylcyclohexanone, via a cyclohexane-1,2-dione (4-
9
Similarly, chemical resolution with many common chiral
acids also failed to give product of high enantiomeric purity
in satisfactory yield. The best chemical resolution was
obtained using one equivalent of dibenzoyl-D-tartaric acid
chlorophenyl)hydrazone intermediate. The reaction, known
as Borsche synthesis, is a special case of the Fischer indole
synthesis.
6
f,g
(
D-DBTA) in methanol, which afforded 2 in 93:7 enantio-
meric ratio, but only 13% yield. While acylation of the
resolved amine (2) with 2-picolyl chloride (6) in the presence
of Hunig’s base provided 1 in 80-90% yield, we had yet to
identify a method that could provide chiral amine 2 in
suitable purity and yield for preclinical development.
The strategy of an enantioselective synthesis utilizing
chiral catalysis was then examined. The direct access of 2
from ketone 3 through reductive amination is the most
obvious methodology for this strategy. While Noyori reduc-
tion has been known for enantioselective reduction of imines,
its utility for the synthesis of a primary amine such as 2 is
limited, due to the low stability as well as the reversibility
of the imine intermediate. Noyori asymmetric hydrogenation
with a chiral Ru(II) catalyst and a 5:2 formic acid/triethyl-
amine azeotropic mixture is typically more enantioselective
on cyclic imines for the preparation of endocyclic chiral
Figure 1. Strategy for synthesis of 1.
Our first synthesis was racemic (Scheme 1). Reductive
amination of 6-chloro-2,3,4,9-tetrahydro-1H-carbazol-1-one
(3) with ammonium acetate and sodium cyanoborohydride
provided racemic amine 2 in 52% yield. Imine formation
from ketone 3 with ammonium acetate proved difficult, and
the reaction required 15 h at 60 °C. A significant amount of
tar was generated, presumably due to the acid-catalyzed
decomposition of the ketone as well as the amine product,
which contributed to the low yield. The racemic amine 2
was resolved by preparative chiral supercritical fluid chro-
1
3a
amines. For the preparation of primary amines such as 2,
Noyori has used an alkyl amine such as benzylamine as the
amino source which could then be cleaved by hydrogenolysis
matography (SFC) using 10% MeOH, 10% CHCl
3
, and 0.5%
1
3a
trifluoroacetic acid (TFA) in carbon dioxide as eluent on a
to provide a primary amine. More recently, Kadyrov and
Daicel AS-H column.1
0,11
While the chiral separation pro-
(11) Conditions for chiral analysis of 2 were as follows. Column: Chiracel OD-
vided enough material to support early studies, the low
H, 4.6 mm × 250 mm; mobile phase: 15:85 i-PrOH/hexanes (0.1 % Et
2
-
NH); flow rate: 1 mL/min; detection: 230 nm; temperature: 25 °C; retention
time: 28 min for (S)-2, 12.3 min for enantiomer (R)-2. The salt form (13)
was analyzed the same way as for free amine 2 after neutralization with
aqueous NaOH and extraction into THF.
(8) Tishler, M.; Gal, G.; Stein, G. A. Org. Synth. 1959, 39, 27.
(9) 6-Chloro-2,3,4,9-tetrahydro-1H-carbazol-1-one (3) was obtained from Ubichem
Research through custom synthesis.
(
10) Racemic amine 2 was resolved on a Berger analytical SFC with an HP1100
diode array detector. The sample was monitored at 230 nm under the
following conditions: 10% methanol, 10% chloroform, 0.5% trifluoroacetic
acid in carbon dioxide with a flow rate of 2 mL/min at 1500 psi, 27 °C on
a Daicel AD-H column (3 cm × 25 cm). When the separated enantiomers
were run on an analytical column Daicel AD-H (4.6 mm × 250 mm) in
(12) Racemic final product 1 was resolved on a Berger analytical SFC with an
HP1100 diode array detector. The sample was monitored at 230 nm under
the following conditions: 30% methanol in carbon dioxide with a total flow
rate of 2 mL/min at 1500 psi, 40 °C on a Daicel AS-H column (4.6 mm ×
250 mm), to give the R-isomer (retention time 5.08 min) and the S-isomer
(retention time 7.45 min).
(13) (a) Uematsu, N.; Fujii, A.; Hashiguchi, S.; Ikariya, T.; Noyori, R. J. Am.
Chem. Soc. 1996, 118, 4916. (b) Kadyrov, R.; Riermeier, T. H. Angew.
Chem., Int. Ed. 2003, 42, 5472.
1
0% methanol, 10% chloroform, 0.5% trifluoroacetic acid in carbon dioxide
with a flow rate of 2 mL/min at 1500 psi, 27 °C, the R-isomer had a retention
time of 5.57 min, and the S-isomer a retention time of 8.13 min.
540
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

Scheme 2
Table 1. Selectivity Observed with Various Chiral
-(Phenyl)ethylamines
1
diastereo ratio
after salt
isolated yield
chiral
diastereo ratio
isolated from ketone 3
entry auxiliary from reduction crystallization product
(%)
1
2
3
10a
10b
10c
96:4
95:5
88:12
>99.5:0.5
>99.5:0.5
99.3:0.7
12a
12b
12c
82
86
70
carbazolone via reductive amination, although the stereose-
1
4a
lectivity was not verified.
We were delighted to find that several chiral (phenyl)-
ethylamines gave high diastereofacial selectivity in the
reductive amination of tetrahydrocarbazolone 3 as shown in
Scheme 3. The reductive amination could be carried out in
one pot with the ketone and reagents mixed in dichlo-
romethane or 1,2-dichloroethane. However, the reactions on
large scale were found to be more reproducible when
performed in two sequential steps: imine formation with
catalytic amount of p-TsOH or HCl in toluene at 105 °C,
followed by reduction of the resultant imine intermediates
co-worker reported enantioselective hydrogen-transfer reduc-
tive amination with ammonium formate and Ru(II)-BINAP
catalysts for preparation of primary amines.1
3b
(11a-c) in a mixture of toluene and ethanol at -30 °C to
We pursued both types of catalyst and achieved modest
success (Scheme 2). The Noyori transfer hydrogenation was
performed on ketone 3 with 2 mol % of Noyori’s Ru(II)
room temperature. As shown in Table 1, the best selectivity
was achieved with (R)-1-(4-methoxyphenyl)ethylamine (10a,
9
6:4), and with (R)-1-(phenyl)ethylamine (10b, 95:5), while
1
3a
catalyst 7 and 6 equiv of ammonium formate in methanol
at 60 °C under nitrogen atmosphere for 18 h. This modified
Noyori condition provided amine 2 in 9:1 enantiomeric ratio
and 60% isolated yield. Alternatively, the transfer hydroge-
nation could be performed on the preformed imine 8 with
the catalyst prepared in situ from (R)-BINAP (9) and
benzeneruthenium(II) chloride dimer.1 Imine 8 was formed
in 7 N ammonia in methanol with catalytic amount of
p-toluenesulfonic acid. This method provided 2 in similar
enantioselectivity (91:9) but of unacceptable chemical purity.
While the possibility of improving the yield from the
Noyori transfer hydrogenation existed, we did not have a
good method to enhance the enantiomeric purity from the
15
(S)-phenylglycinol (10c) gave a poorer ratio (88:12). In all
cases, the crude amino product was treated with HCl and
isolated as a HCl salt. The formation and crystallization of
the HCl salt enhanced the diastereomeric purity of the
isolated product to over 99% as determined by analysis by
15
HPLC. The undesired diastereo isomer was essentially
3b
removed in just one crystallization. The reduction of the
methoxy compound (11a) was selected for scale-up because
of the more facile cleavage of the chiral auxiliary (vide infra).
On scale-up, 12a was isolated in 82% yield and 99.6%
diastereomeric purity on a 300-gal scale. The yield ap-
proached 90% in smaller fixed equipment.
With the preparation of the “protected” chiral amines
(12a-c) achieved in high yield and excellent selectivity,
deprotection and amide formation were the two remaining
steps. Cleavage of the chiral auxiliaries to give the desired
chiral amine 2 proved difficult. Hydrogenolysis with various
metal catalysts (Pd, Ru, and Rh) is the most common method
80% ee delivered by the chiral reduction to a level acceptable
for pharmaceutical development (>98% ee). The cost of the
chiral Ru(II) catalysts was also an issue for the industrial
production we were pursuing. Therefore, our efforts shifted
to an approach based on asymmetric reduction directed with
a chiral auxiliary. For this approach, we focused on chiral
benzyl amines that are known surrogates for the amino group
1
6
for N-debenzylation, but it led to loss of the chloro
substituent as well as decomposition of the tetrahydrocar-
bazole moiety in all cases (12a-c). Three equivalents of
14
and available on scale at low cost. Li e´ geois and co-workers
used (S)- and (R)-R-methylbenzylamines in their efforts
towards synthesis of pirlindole from a 6-methyl tetrahydro-
AlCl
3
in dichloromethane worked to some extent in benzyl
cleavage of the benzyl groups of 12a and 12b. However,
Scheme 3
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
541

Scheme 4
the low stability of the tetrahydrocarbazole moiety in the
presence of the large amount of strong acids generated in
the difficult acidic workup led to only about 30% isolated
yield of the HCl salt of 2. For the hydroxyl compound 12c,
we found that the phenylglycinol moiety could be removed
as described herein has been scaled up in pilot-plant reactors
and has produced multikilograms of 1. Assignment of the
absolute stereochemistry of 1 was accomplished by X-ray
analysis of a single crystal of 1 as shown in Figure 2.
4 3 3
with NaIO . Eventually, we found that BCl and BBr gave
the cleanest N-debenzylation for 12a and 12b. The boron
reagents worked particularly well for 12a because of the
stabilization effect on the purported benzyl cationic inter-
mediate from the p-methoxy group. Compound 12a was
chosen for scale-up in the pilot plant (Scheme 4). The
N-debenzylation progressed smoothly with 2-3 equiv of
3
BCl in dichloromethane. However, isolation of free base 2
could not be achieved consistently with satisfactory purity
and yield. After extensive screening of various acids for the
salt formation and crystallization, we were rewarded with a
pleasant surprise. The only crystalline salt with acceptable
isolated purity and yield turned out to be the 1:1 salt (13)
from 2 and 2-picolinic acid; the same acid that was to be
incorporated into the target 1 in the last step. In scale-up to
Figure 2. ORTEP view of X-ray structure of 1.
200 gallons, salt 13 was isolated in 81% yield and 99.2%
ee. With a second crop, the total yield was as high as 92%.
While typical acylation conditions such as O-benzotriazole
activation and acid chloride worked for the formation of the
amide from 13, the most practical method utilized 1-propyl-
phosphonic acid cyclic anhydride (T3P) and Hunig’s base
Conclusion
An efficient asymmetric synthesis of N-[(1R)-6-chloro-
2
,3,4,9-tetrahydro-1H-carbazol-1-yl]-2-pyridinecarboxam-
ide has been achieved. Three generations of synthesis were
explored: a racemic synthesis followed by a resolution
through chiral chromatography or crystallization of the salt
of the amine with a chiral acid, an enantioselective asym-
metric synthesis through a reductive amination via chiral
transfer hydrogenation catalyzed by two Noyori chiral Ru
17
in dichloromethane (Scheme 4). The carboxylic acid,
activated via a purported mixed anhydride with a 1-propy-
lphosphonic acid moiety, reacted with the amine (2) cleanly
and nearly instantaneously with ice cooling. The target active
pharmaceutical ingredient (API) 1 was isolated as a crystal-
(II) complexes, and a chiral (phenyl)ethylamine-directed
18
line solid in up to 87% yield in g99.5% ee. The sequence
diastereo or facial selective synthesis. The latter was the most
efficient and most selective. The synthesis also featured
isolation of the 2-picolinic acid salt of (1R)-6-chloro-2,3,4,9-
tetrahydro-1H-carbazol-1-amine and application of 1-pro-
pylphosphonic acid cyclic anhydride (T3P) for the convenient
amide formation from the two components of the salt. This
practical stereoselective synthesis has been scaled up in 200-
gallon reactors for delivery of multikilograms of the target
compound in over 99.5% enantiomeric purity.
(
14) (a) Tullio, P. d.; Felikidis, A.; Pirotte, B.; Li e´ geois, J.-F.; Stachow, M.;
Delarge, J. HelV. Chim. Acta 1998, 81, 539. (b) Gutman, A. L.; Etinger,
M.; Nisnevich, G.; Polyak, F. Tetrahedron: Asymmetry 1998, 9, 4369.
15) HPLC conditions for analysis of diastereomers 12a-c with HPLC were as
follows: Column: Luna C18(2) 50 mm × 2 mm, 3 µm; mobile phase A:
(
2
H O (0.05% TFA), mobile phase B: MeCN (0.05% TFA); gradient: 0-95%
B over 8 min; detection: 220 nm; temperature: 40 °C; retention time: 4.45
min for 12a and 4.55 min for its diastereomer; 4.38 min for 12b and 4.49
min for its diastereomer; 4.21 min for 12c and 4.32 min for its diastereomer.
16) (a) Rylander, P. N. Catalytic Hydrogenation in Organic Synthesis; Academic
Press: London, 1979. (b) Frahm, A. W.; Knupp, G. Tetrahedron Lett. 1981,
(
2
2, 2633. (c) Speckenbach, B.; Pisel, P.; Frahm, A. W. Synthesis 1997,
Experimental Section
1
325.
(
17) (a) Wissmann, H.; Kleiner, H., J. Angew. Chem., Int. Ed. Engl. 1980, 19,
(1R)-6-Chloro-2,3,4,9-tetrahydro-1H-carbazol-1-am-
ine (2). 1. Racemic Synthesis Followed by Resolution by
Chiral Column. Racemic Amine. To a solution of 6-chloro-
2,3,4,9-tetrahydro-1H-carbazol-1-one (3) (500 mg, 2.3 mmol)
and ammonium acetate (1.8 g, 23 mmol) in methanol (9 mL)
was added sodium cyanoborohydride (720 mg, 11.5 mmol).
1
33. (b) Helal, C. J.; Kang, Z.; Lucas, J. C.; Bohall, B. R. Org. Lett. 2004,
6
, 1853.
(
18) HPLC conditions for chiral analysis of 1 were as follows. Column: Chiracel
OD-H, 4.6 mm × 250 mm; mobile phase: 10:90 i-PrOH/hexanes (0.1%
2
Et NH); flow rate: 1 mL/min; detection: 230 nm; temperature: 25 °C;
retention time: 9.3 min for enantiomer (S)-1, 11.2 min for enantiomer (R)-
1
.
542
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

After heating at 60 °C for 15 h, the mixture was cooled and
treated with concentrated hydrochloric acid until pH ) 1.
The organics were removed under reduced pressure, and the
resulting precipitate was collected by filtration, dissolved in
For the salt formation in the same reactor, the crude amine
prepared as above was diluted with 34.4 L of methanol and
51.6 mL of toluene. After being cooled to -5 °C, the mixture
was treated with 7.05 kg (101 mol) of concentrated HCl over
30 min and stirred for additional 30 min. The solid salt was
filtered, and the filter cake was washed with 6.9 L of methyl
tert-butyl ether (MTBE). The filter cake was dried at 60 °C
under vacuum to provide 25.1 kg (82% overall yield from
ketone 3) of HCl salt 12a as a white crystalline solid:
30 mL of ethyl acetate and 5 mL of methanol, and washed
with 20 mL of saturated aqueous sodium carbonate. The
phases were separated, and the organic phase was concen-
trated to yield racemic 6-chloro-2,3,4,9-tetrahydro-1H-car-
bazol-1-amine (2, racemic) (260 mg, 52% yield) as a light-
15
22
1
diastereomeric ratio 99.6:0.4; [R] +89.9 (c 0.55, MeOH);
brown solid. H NMR (DMSO-d
6
) δ 10.90 (s, 1H), 7.34 (m,
H), 7.27 (d, 1H), 6.97 (dd, 1H), 3.90 (t, 1H), 2.54 (m, 2H),
.04-1.89 (m, 2H), 1.66 (m, 1H), 1.50 (m, 1H); MS m/z
D
1
H NMR (DMSO-d
6
) δ 11.9 (s, 1H), 10.0 (m, 1H), 9.35 (m,
1
2
2
1
H), 7.70 (d, J ) 8.7 Hz, 2H), 7.52 (s, 1H), 7.40 (d, J ) 8.7
Hz, 1H), 7.13 (d, J ) 8.6 Hz, 1H), 7.00 (d, J ) 8.7 Hz,
2H), 4.69 (m, 1H), 4.54 (m, 1H), 3.75 (s, 3H), 2.62 (m, 2H),
21 (M + 1); HRMS (ESI): calcd for C12
H
11ClN (M -
+
NH
2
) 204.0580, found 204.0570.
2
.14 (m, 1H), 2.01 (m, 2H), 1.76 (m, 1H), 1.67 (d, J ) 8.6
2
. EnantioselectiVe Synthesis through Chiral Transfer
Hydrogenation. A mixture of 220 mg (1.00 mmol) of ketone
, 380 mg (6.00 mmol) of ammonium formate, and 12 mg
1
3
Hz, 3H); C NMR (CDCl
3
) δ 159.5, 134.3, 130.1, 129.9,
1
29.6, 127.1, 123.4, 122.1, 117.8, 114.0, 113.1, 112.9, 55.2,
3
13a
55.0, 49.1, 25.7, 20.1, 19.7, 18.8. HRMS calcd for C21
H
24
-
(0.020 mmol) of Noyori’s catalyst in 5 mL of methanol
+
ClN
2
O (M + H) 355.1577, found, 355.1577. Anal. Calcd
was heated to 60 °C under nitrogen atmosphere for 18 h.
The mixture was cooled to ambient temperature and quenched
with 20 mL of 2 M HCl in water. The mixture was washed
for C21
H
23ClN
2
O‚HCl: C, 64.45; H, 6.18; N, 7.16; Cl, 18.12.
Found: C 64.26; H, 6.12; N, 7.15; Cl, 18.40.
1R)-1-(5-Chloro-3-propyl-1H-indol-2-yl)-N-[(1R)-1-
(
with 10 mL of CH
phase was made basic with 2 M NaOH in water and extracted
with 15 mL of CH Cl twice. The combined organic layers
were dried over anhydrous Na SO and evaporated under
vacuum to give 132 mg (60%) of product 2 as a gray solid:
2 2
Cl . Layers were separated. The aqueous
phenylethyl]ethanamine (12b). To a 20-L reactor with a
condenser and built-in Dean-Stark trap were successively
added 1.00 kg (4.55 mol) of 6-chloro-2,3,4,9-tetrahydro-1H-
carbazol-1-one (3), 16 L of toluene, 552 g (4.55 mol) of
2
2
2
4
(+)-R- methylbenzylamine (10b), and 60.5 g (318 mmol)
11
enantiomeric ratio 9:1; purity by HPLC 98%. For the final
synthetic route, amine 2 was isolated and characterized as a
salt with 2-picolinic acid. See the procedure for compound
of p-TsOH‚H O. The mixture was heated at reflux for 12 h.
2
The reaction mixture was partially concentrated to 4 L under
vacuum. The crude imine in toluene was diluted with 12 L
of EtOH, cooled to -17 °C. To the mixture (which might
be a slurry at this point) was added 1.37 L (2.74 mol) of 2
13.
(1R)-1-(5-Chloro-3-propyl-1H-indol-2-yl)-N-{(1R)-1-[4-
(methoxy)phenyl]ethyl}ethanamine (12a). To a 300-gal
M NaBH
4
in triglyme over 1 h between -17 °C to -20 °C.
. The
reactor with a condenser and built-in Dean-Stark trap were
successively added 17.2 kg (78.3 mol) of 6-chloro-2,3,4,9-
tetrahydro-1H-carbazol-1-one (3), 206 L of toluene, 12.4 kg
A solution formed after the addition of the NaBH
4
mixture was stirred at -15 °C for 10 h. The reaction mixture
was sampled for analysis by HPLC which showed a
(82.0 mol) of (R)-(+)-1-(4-methoxyphenylethyl)amine (10a),
diastereomeric ratio of 95:5 favoring the desired diastereomer
and 276 g (2.78 mol) of concentrated HCl. The reaction was
brought to reflux and was stirred for 10 h. After being cooled
to room temperature, the mixture containing imine interme-
diate 11a was drained to an addition vessel. The reactor was
rinsed with 21.5 L of toluene which was added to the addition
vessel.
The reactor was then charged with 172 L of ethanol and
cooled to -30 °C, followed by addition of 2.92 kg (77.3
mol) of sodium borohydride in three portions. The crude
imine from the addition vessel was added over about 4 h
with temperature maintained at about -30 °C. After being
stirred for 12 h, the reaction was warmed to room temper-
ature over 1 h. The reaction was then treated with 8.6 L of
acetone and stirred for 30 min. The volume of the reaction
mixture was reduced to about 70 L with vacuum distillation,
followed by addition of 258 L of ethyl acetate and 103 L of
water. The layers were separated. The organic layer was
successively washed with 85 L of water and 85 L of 10 wt
15
1
2b. The reaction mixture was warmed to room temperature
over 1 h, and 0.5 L of acetone was added over 30 min. The
mixture was concentrated to 4 L, followed by addition of
1
3
4 L of ethyl acetate and 5 L of 4% NaHCO aqueous
solution. Layers were separated, and the organic layer was
successively washed with 4 L of water and 4 L of 10% brine
and concentrated to 8 L under vacuum.
After being diluted with 5 L of ethyl acetate and 2 L of
MeOH, 1.50 L (6.00 mol) of 4 N HCl in 1,4-dioxane was
added over 20 min. The resultant slurry was stirred for 1 h
at ambient temperature and filtered. The filter cake was
washed with 2 L of EtOH and dried under vacuum at 60 °C
for 24 h to provide 1.41 kg (86% overall yield from ketone
3
) of HCl salt 12b as a white crystalline solid: diastereomeric
1
5
22
1
ratio 100:0; [R]
D
+149 (c 1.23, MeOH); H NMR
(DMSO-d ) δ 11.7 (s, 1H), 9.93 (m, 1H), 9.34 (m, 1H), 7.82
6
(d, J ) 8.7 Hz, 2H), 7.50 (s, 1H), 7.47 (m, 3H), 7.18 (d, J
) 8.7 Hz, 2H), 4.77 (m, 1H), 4.61 (m, 1H), 2.67 (m, 2H),
2.53 (m, 3H), 2.20 (m, 1H), 2.02 (m, 2H), 1.85 (m, 1H),
%
brine. Analysis by HPLC showed the diastereomeric ratio
of 96:4 for free base of 12a and its diastereomer at this
point.15
1.73 (d, J ) 8.6 Hz, 3H); HRMS (ESI): calcd for C20
H
22
-
+
ClN
2
(M + H) 325.1472, found 325.1474. Anal. Calcd for
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
543

C
20
H
21ClN
2
‚HCl: C, 66.49; H, 6.14; N, 7.75; Cl, 19.62.
Found: C, 66.21; H, 6.18; N, 7.70; Cl, 19.54.
2S)-2-{[(1R)-1-(5-Chloro-3-propyl-1H-indol-2-yl)ethyl]-
organic layers were washed with 135 L of 10 wt % brine to
give crude free amine 2 as a solution in dichlorometh-
ane.
(
amino}-2-phenylethanol (12c). To a three-neck round-
bottom flask equipped with a condenser and Dean-Stark
trap were added 2.00 g (9.02 mmol) of 6-chloro-2,3,4,9-
tetrahydro-1H-carbazol-1-one (3), 20 mL of toluene, 1.40 g
To the crude amine as described above was added 4.73
kg (38.3 mol) of 2-picolinic acid. The solution was concen-
trated to about 95 L via atmospheric distillation. After being
diluted with 95 L of isopropanol, the mixture was further
concentrated to about 95 L via atmospheric distillation. The
mixture was cooled to 10 °C and stirred for 1 h. The mixture
was filtered, and the filter cake was washed with 27 L of
(9.47 mmol) of (S)-phenylglycinol (10c), and 90.0 mg (0.500
mmol) of p-toluenesulfonic acid. The reaction was brought
to reflux and was stirred for 24 h. After being cooled to room
temperature, the mixture containing imine intermediate 11c
was concentrated to 8 mL.
A second round-bottom flask was charged with 10 mL
of methanol and cooled to -45 °C, followed by addition of
2
:1 mixture of isopropyl alcohol/dichloromethane. The filter
cake was dried at 60 °C under vacuum to provide 9.6 kg
81%) of the 2-picolinic acid salt 13 as a crystalline solid:
(
1
1
22
enantiomeric ratio 99.6:0.4; [R]
D
+159 (c 0.53, MeOH);
385 mg (9.47 mmol) of sodium borohydride. The crude imine
1
H NMR (DMSO-d ): δ 8.61 (s, 1H), 8.00 (d, J ) 7.8 Hz,
6
from the addition vessel was added with temperature
maintained at approximately -45 °C. Once addition of the
imine was complete the reaction was warmed to -15 °C.
After being stirred for 2 h, the reaction was warmed to room
temperature. The reaction was then treated with 1 mL of
acetone and stirred for 30 min. The reaction mixture was
concentrated to about 8 mL under vacuum, followed by
addition of 28 mL of ethyl acetate and 5 mL of 5% sodium
chloride in water. The layers were separated, and the organic
layer was washed with 5 mL of 10 wt % brine. Analysis by
HPLC showed the diastereomeric ratio of 88:12 at this
point.15
For the salt formation in the same reactor, the free base
of 12c as described above was diluted with 4 mL of methanol
and 10 mL of ethyl acetate. The mixture was treated with 3
mL (11.8 mmol) of HCl in 1,4-dioxane over 5 min and stirred
for additional 20 min. The solid was filtered, and the filter
cake was washed with 4 mL of methyl tert-butyl ether
1
7
2
H), 7.85 (m, 1H), 7.46 (s, 1H), 7.40 (d, H ) 4.3 Hz, 1H),
.09 (d J ) 4.3 Hz, 1H), 4.43 (s, 1H), 2.62 (m, 2H),
.14 (m, 1H), 1.98 (m, 1H), 1.79 (m, 2H). HRMS (ESI):
+
calcd for C12
H
11ClN (M - NH
2
)
204.0580, found
2
04.0576.
N-[(1R)-6-Chloro-2,3,4,9-tetrahydro-1H-carbazol-1-yl]-
-pyridinecarboxamide (1). To a 200-gal reactor was
2
successively added 12.9 kg (37.5 mol) of salt 13, 90 L of
dichloromethane, and 16.0 kg (7.68 mol) of Hunig’s base.
To ensure a complete reaction, 464 g (3.76 mol, 0.1 equiv
to 13) of 2-picolinic acid was added. After being cooled to
0 °C, the slurry was treated with 33.4 kg (52.5 mol) of 50
wt % solution of 1-propylphosphonic acid cyclic anhydride
(T3P) in EtOAc over 30 min while maintaining the temper-
ature at 0-10 °C. The off-white slurry became a clear brown
solution and was stirred at 0 °C for about 2 h. The reaction
was quenched by addition of 90 L of water at below 25 °C
and stirred for 30 min. Layers were separated, and the
aqueous layer was extracted with 65 L of dichloromethane.
The combined organic layers were successively washed with
(MTBE) followed by 4 mL of ethyl acetate. The filter cake
was dried at 60 °C under vacuum to provide 2.46 g (70%
overall yield from ketone 3) of HCl salt 12c as a crystalline
solid: diastereomeric ratio 99.3:0.7.1 H NMR (DMSO-d
5 1
6
)
9
0 L of water and 90 L of 15 wt % brine.
δ 11.6 (s, 1H), 9.83 (m, 1H), 9.49 (m, 1H), 7.74 (d, J ) 8.7
Hz, 2H), 7.50 (s, 1H), 7.43 (m, 3H), 7.14 (d, J ) 8.7 Hz,
The organic layer was concentrated to about 52 L via
atmospheric distillation to azeotropically reduce water content
to below 0.21 wt % by Karl Fisher analysis. This distillation
was repeated if necessary in order to achieve the target water
level. The solution was diluted with 65 L of ethanol and
concentrated to about 65 L under vacuum. The resultant
slurry was treated with 206 L of ethanol to make a solution
at 76-80 °C. A hot clarification filtration was carried out
on the solution containing 1. The filtrate was cooled to 20-
2
1
2
H), 5.59 (br s, 1H), 4.72 (m, 2H), 4.11 (m, 1H), 3.87 (m,
H), 3.50 (br s, 1H), 2.64 (m, 2H), 2.30 (m, 1H), 2.01 (m,
H), 1.75 (m, 1H). HRMS (ESI): calcd for C20
H22ClN
2
O
+
(M + H) 341.1421, found 341.1418.
(
1R)-6-Chloro-2,3,4,9-tetrahydro-1H-carbazol-1-am-
monium 2-picolinate (13). To a 200-gal reactor were
successively added 13.5 kg (34.5 mol) of 12a and 67.5 L of
dichloromethane. With reaction temperature maintained at
2
5 °C and treated with 13 L of water over 1 h with stirring.
3
about 25 °C, 86.3 L (86.3 mol) of 1 M BCl in dichlo-
The mixture was cooled to 0 °C and stirred for 2 h. The
resultant white slurry was filtered. The filter cake was washed
with 65 L of 1:1 mixture of ethanol-water, dried at 80 °C
under vacuum to provide 9.7 kg (79%) of target compound
romethane was added over about 30 min. After being stirred
at 25 °C for 3 h, the reaction was cooled to 10 °C and stirred
overnight. The reaction was quenched with 135 kg (10
weights of 12a) of 20 wt % aqueous KOH over about 1 h
such that the first two weights were added with vigorous
stirring over 30 min to maintain the temperature at about 10
1
[R]
as a white crystalline solid: enantiomeric ratio >99.8:0.2;18
1
D
) + 192 (c 1.16, CH
2
Cl
2
); H NMR (DMSO-d
6
): δ
°
C. The mixture was warmed to 25 °C. Layers were
10.93 (s, 1H), 8.78 (d, 1H), 8.61 (m, 1H), 8.12 (m, 1H),
8.02 (m, 1H), 7.61 (m, 1H), 7.43 (m, 1H), 7.26 (d, 1H), 7.01
(dd, 1H), 5.34 (m, 1H), 2.64 (m, 2H), 2.00 (m, 3H), 1.82
separated, and the aqueous layer was extracted with 135 L
of 9:1 mixture of dichloromethane/methanol. The combined
544
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

(
1
1
m, 1H); 13C NMR (DMSO-d
6
): δ 164.1, 150.6, 149.1,
38.4, 136.4, 135.3, 128.5, 127.2, 123.7, 122.7, 121.5, 117.8,
Acknowledgment
We thank Mr. David Black for conducting HRMS
experiments.
13.4, 111.2, 44.0, 30.8, 21.5, 21.0. HRMS (ESI): calcd for
+
C
18
H
17ClN
Calcd for C18
N, 12.72. Found: C, 65.67; H, 4.91; N, 12.66.
3
O (M + H) 326.1060, found 326.1061. Anal.
Received for review October 27, 2006.
OP060223V
3 2
H16ClN O with 1/4 H O: C, 65.45; H, 5.04;
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
545