Large-scale asymmetric synthesis of the 3,6,7,8-tetrahydrochromeno[7,8-d] imidazole BYK 405879: A promising candidate for the treatment of acid-related diseases

Article abstract of DOI:10.1021/op800177x

A process for the synthesis of the potassium-competitive acid blocker BYK 405879 (8) was established based on the approach used in medicinal chemistry (asymmetric hydrogenation of prochiral ketone 15 and Mitsunobu cyclization of the resulting alcohol 34). Several critical reaction steps were optimized. The synthesis of prochiral ketones was accomplished using ethyl 3-(2-methylphenyl)- 3-oxopropanoate instead of 1-[1-(2-methylphenyl)vinyl]pyrrolidine, a reagent that was difficult to prepare and possesses limited shelf life. The catalyst loading of the asymmetric hydrogenation step was reduced significantly from a S/C ratio of 100:1 to a S/C ratio of 2500:1 by benzyl protection of ketone 15. After the Mitsunobu cyclization, the removal of byproduct was easily accomplished through acid-base extraction, and pure BYK 405879 (8) was then obtained by means of crystallization in the presence of succinic acid.

Full text of DOI:10.1021/op800177x

Organic Process Research & Development 2008, 12, 1170–1182
Large-Scale Asymmetric Synthesis of the
3,6,7,8-Tetrahydrochromeno[7,8-d]imidazole BYK 405879: A Promising Candidate for
the Treatment of Acid-Related Diseases
Andreas M. Palmer,*,† Matthias Webel,*,‡ Christian Scheufler,‡ Dieter Haag,§ and Bernd Mu¨ller§
NYCOMED GmbH, Department of Medicinal Chemistry and NYCOMED GmbH, Department of Process Chemistry Research,
Byk-Gulden-Strasse 2, D-78467 Konstanz, Germany, and NYCOMED GmbH, Department of Process Chemistry DeVelopment,
Robert-Bosch-Strasse 8, D-78224 Singen, Germany
Abstract:
of GERD.4 Irreversible inhibitors of the H+/K+-ATPase (PPIs),
e.g. lansoprazole, omeprazole, or pantoprazole have already
been available for some time and are a popular choice for the
therapy of GERD. Due to its systemic character, the PPI
treatment marked a major breakthrough compared to the
classical therapy with antacids. Nevertheless, PPIs exhibit
substantial interpatient variability and often fail to provide a
complete cure. Hence, several pharmaceutical companies are
engaged in the development of potassium-competitive acid
blockers (P-CABs). Due to their different mode of action
(reversible inhibition of the gastric proton pump enzyme),
P-CABs might be able to overcome some limitations observed
during the PPI treatment of GERD.5-8
In the course of our P-CAB lead optimization program, we
identified several compounds belonging to the classes of 7H-
8,9-dihydropyrano[2,3-c]imidazo[1,2-a]pyridines 3 (e.g., BYK
311319, 4) and 3,6,7,8-tetrahydrochromeno[7,8-d]imidazoles 7
(e.g., BYK 405879, 8) that are potent reversible inhibitors of
the gastric proton pump enzyme and possess promising
pharmacological properties.9,10 Consequently, we designed an
asymmetric synthesis that permitted the ready preparation of
multigram quantities of target compounds 3 and 7 (Scheme
1).10,11
A process for the synthesis of the potassium-competitive acid
blocker BYK 405879 (8) was established based on the approach
used in medicinal chemistry (asymmetric hydrogenation of prochiral
ketone 15 and Mitsunobu cyclization of the resulting alcohol 34).
Several critical reaction steps were optimized. The synthesis of
prochiral ketones was accomplished using ethyl 3-(2-methylphe-
nyl)-3-oxopropanoate instead of 1-[1-(2-methylphenyl)vinyl]pyr-
rolidine, a reagent that was difficult to prepare and possesses
limited shelf life. The catalyst loading of the asymmetric hydro-
genation step was reduced significantly from a S/C ratio of 100:1
to a S/C ratio of 2500:1 by benzyl protection of ketone 15. After
the Mitsunobu cyclization, the removal of byproduct was easily
accomplished through acid-base extraction, and pure BYK
405879 (8) was then obtained by means of crystallization in the
presence of succinic acid.
1. Introduction
Acid-related diseases, such as gastroesophageal reflux disease
(GERD), have a high prevalence and a strong impact on the
quality of life of the affected patients.1,2 Moreover, they
represent a significant economical burden. Whereas the most
common symptom of GERD is frequent heartburn, more serious
diseases including esophagitis and/or esophageal cancer may
result if the disease remains untreated. Although the pathogen-
esis of GERD is complicated, the reflux of gastric acid into the
esophagus was identified as one major factor.3 The inhibition
of the gastric proton pump enzyme (H+/K+-ATPase) located
in the parietal cell provides a valuable approach for the treatment
In both classes, the same retrosynthetic strategy was applied
comprising the asymmetric catalytic hydrogenation of ketones
1 and 5 and subsequent Mitsunobu cyclization of the resulting
(4) Shin, J. M.; Vagin, O.; Munson, K.; Kidd, M.; Modlin, I. M.; Sachs,
G. Cell. Mol. Life Sci. 2008, 65, 264–281.
(5) Parsons, M. E.; Keeling, D. J. Expert Opin. InVest. Drugs 2005, 14,
411–421.
(6) Fass, R.; Shapiro, M.; Dekel, R.; Sewell, J. Aliment. Pharmacol. Ther.
2005, 22, 79–94.
(7) Vakil, N. Aliment. Pharmacol. Ther. 2004, 19, 1041–1049.
(8) Andersson, K.; Carlsson, E. Pharmacol. Ther. 2005, 108, 294–307.
(9) (a) Palmer, A. M.; Grobbel, B.; Jecke, C.; Brehm, C.; Zimmermann,
P. J.; Buhr, W.; Feth, M. P.; Simon, W.-A.; Kromer, W. J. Med. Chem.
2007, 50, 6240–6264. (b) Buhr, W.; Chiesa, M. V.; Zimmermann,
P. J.; Brehm, C.; Simon, W.-A.; Kromer, W.; Postius, S.; Palmer, A.
Patent application WO 2005/058325, 2005.
(10) (a) Palmer, A. M.; Chiesa, V.; Holst, H. C.; Le Paih, J.; Zanotti-Gerosa,
A.; Nettekoven, U. Tetrahedron: Asymmetry 2008, 19, 2102-2110.
(b) Buhr, W.; Chiesa, M. V.; Zimmermann, P. J.; Brehm, C.; Palmer,
A.; Postius, S.; Kromer, W.; Simon, W.-A.; Senn-Bilfinger, J. Patent
Application WO 2004/087701, 2004. (c) Chiesa, M. V.; Palmer, A.;
Buhr, W.; Zimmermann, P. J.; Brehm, C.; Simon, W.-A.; Postius, S.;
Kromer, W.; Zanotti-Gerosa, A. Patent application WO 2006/136552,
2006.
* To whom correspondence should be addressed. Telephone: +49 7531 84
4783 (A.M.P.); +49 7531 84 4563 (M.W.). Fax: +49 7531 849 2087 (A.M.P.);
+49 7531 849 4563 (M.W.). E-mail: andreas.palmer@nycomed.com;
matthias.webel@nycomed.com.
† NYCOMED GmbH, Department of Medicinal Chemistry, Konstanz,
Germany.
‡ NYCOMED GmbH, Department of Process Chemistry Research, Konstanz,
Germany.
§ NYCOMED GmbH, Department of Process Chemistry Development,
Singen, Germany.
(1) Modlin, I. M.; Sachs, G. Acid Related Diseases: Biology and
Treatment; Schnetztor-Verlag: Konstanz, Germany, 1998.
(2) Heartburn, hiatal hernia, and gastroesophageal reflux disease (GERD);
NIH: Washington, DC, 2003; Publication No. 03-0882; http://
digestive.niddk.nih.gov/ddiseases/pubs/gerd/index.htm.
(3) Vesper, B. J.; Altman, K. W.; Elseth, K. M.; Haines, G. K., III;
Pavlova, S. I.; Tao, L.; Tarjan, G.; Radosevich, J. A. Chem. Med.
Chem. 2008, 3, 552–559.
(11) (a) Palmer, A. M.; Nedden, H.; Zanotti-Gerosa, A. Tetrahedron:
Asymmetry 2008, 19, 1310–1327. (b) Palmer, A.; Buhr, W.; Zimmer-
mann, P. J.; Brehm, C.; Chiesa, M. V.; Zanotti-Gerosa, A. Patent
application WO 2007/141253, 2007.
1170
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development
10.1021/op800177x CCC: $40.75
2008 American Chemical Society
Published on Web 09/25/2008

Scheme 1a
a Step (i): asymmetric hydrogenation; step (ii): Mitsunobu cyclization.
Figure 2. Structure of byproduct 10 formed during catalytic
asymmetric hydrogenation of ketone 5.
giving the corresponding chiral alcohols 2 in excellent yields
and optical purities (e.g., R′ ) R′′ ) Me, Ar ) Ph: 1.2 equiv
of t-BuOK, catalyst 9, S/C ) 1000:1, 2-PrOH, t-BuOH, H2O,
80 bar H2, 65 °C, 22 h, 93% yield, 98% ee or 1.1 equiv of
t-BuOK, catalyst 9, S/C ) 5000:1, 2-PrOH, t-BuOH, H2O, 25
bar H2, 75 °C, 15 h, 75% yield, 97% ee).11 In contrast, in the
benzimidazole class, higher amounts of hydrogenation catalyst
were required to ensure a smooth transformation of ketones 5
to diols 6 (e.g., R′ ) R′′ ) Me, Ar ) Ph: 1.5 equiv of t-BuOK,
catalyst 9, S/C ) 310:1, 2-PrOH, t-BuOH, H2O, 80 bar H2, 65
°C, 23 h, 70%, 98% ee).10 Moreover, under certain reaction
conditions (high S/C ratio, presence of a large excess of a base),
a competing reaction was observed.10 The formation of side
products 10 (Figure 2) is explained by the fact that the presence
of an excess of a base not only increased the rate of the asym-
metric hydrogenation reaction, but also resulted in the formation
of a reactive enolate species that attacked the carboxamide
moiety of the respective ketone 5. A comparable side reaction
has never been observed in the class of imidazo[1,2-a]pyridines.
The Mitsunobu cyclization of the diols 2 and 6 was performed
under standard conditions using triphenylphosphine and DIAD
and the optical purity of the target compounds 3 and 7 always
reflected the enantiopurity of the respective alcohols 2 and 6.10,11
During the in-depth profiling of P-CABs belonging to the
structural classes 3 and 7, the tetrahydrochromeno[7,8-d]imi-
dazole BYK 405879 (8) emerged as a promising candidate for
clinical studies and several kilograms of material were required
for the short-term. It was therefore decided to use the synthetic
strategy applied in medicinal chemistry as an initial basis for
Figure 1. Structure of the hydrogenation catalyst RuCl2[(S)-
Xyl-P-Phos][(S)-DAIPEN] 9.
diols 2 and 6 affording P-CABs 3 and 7. In collaboration with
Johnson Matthey, Catalysis and Chiral Technologies, we
developed for the hydrogenation reaction the previously un-
known catalyst RuCl2[(S)-Xyl-P-Phos][(S)-DAIPEN] 9 (Figure
1), which belongs to the class of Noyori catalysts of the general
formula RuCl2[PP][NN].10-14 The RuCl2[PP][NN]-catalyzed
hydrogenation of ketones is typically performed in the presence
of catalytic amounts of a base using 2-propanol as a solvent.15
However, due to the presence of the acidic phenol group, an
excess of the base was required for the reduction of ketones 1
and 5 (one equivalent of base to deprotonate the phenol moiety
and a small excess to activate the hydrogenation catalyst).10,11
Despite the similarity of ketones 1 and 5, the efficiency of the
hydrogenation reaction varied considerably between the two
structural classes. In the imidazo[1,2-a]pyridine class, the
hydrogenation of ketones 1 was feasible at high S/C ratios
(12) Xyl-P-Phos: 2,2′,6,6′-tetramethoxy-4,4′-bis[di(3,5-dimethylphenyl)phos-
phino]-3,3′-bipyridinyl; DAIPEN: 1,1-di(4-anisyl)-2-isopropyl-1,2-
ethylenediamine.
(13) (a) Wu, J.; Chen, H.; Kwok, W. H.; Lam, K. H.; Zhou, Z. Y.; Yeung,
C. H.; Chan, A. S. C. Tetrahedron Lett. 2002, 43, 1539–1543. (b)
Wu, J.; Chen, H.; Kwok, W.; Guo, R.; Zhou, Z.; Yeung, C.; Chan,
A. S. C. J. Org. Chem. 2002, 67, 7908–7910. (c) Wu, J.; Ji, J.-X.;
Guo, R.; Yeung, C.-H.; Chan, A. S. C. Chem. Eur. J. 2003, 9, 2963–
2968.
(14) Johnson Matthey, Catalysis and Chiral Technologies, 28 Cambridge
Science Park, Milton Road, Cambridge, CB4 0FP, United Kingdom.
(15) Noyori, R.; Ohkuma, T. Angew. Chem., Int. Ed. 2001, 40, 40–73.
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1171

Scheme 2
Scheme 3a
a Reagents and conditions: (i) benzyl chloride, K2CO3, NaI, EtOH, 55-78 °C, 5 h; (ii) NBS, ethyl acetate, 75-85 °C, 3 h, Σ68%; (iii) Ac2O, MsOH, 120 °C, 1-1.5
h, 82%; (iv) FeCl3, charcoal, hydrazine hydrate, MeOH, 70-75 °C, 18 h, 82%; (v) formaldehyde, NaBH3CN, MeOH/AcOH, rt, 2 h, 84%; (vi) POCl3, 75 °C, 2 h, 86%;
(vii) Pd(OAc)2, PPh3, Me2NH (2 M in THF), DMAP, 6 bar CO, DMF, 130 °C, 18 h, 95%; (viii) Pd/C, H2, MeOH, 50 °C, 3-9 h, >78%.
process development and to optimize the respective reaction
steps. At the beginning of the upscale work, the following topics
were identified as major issues:
2. Results and Discussion
2.1. Access to Building Block 11. Although the approach
to benzimidazole 11 pursued in medicinal chemistry consisted
of a rather long linear sequence comprising of 8 steps and
contained at least three critical steps, it was decided to use this
reaction pathway to attain a prompt access to sufficient
quantities of intermediate 11 urgently needed for the production
of API (8) to support the early development phase (Scheme
3).18 Accordingly, the synthesis was carefully scaled up taking
into account all necessary security measures.
Benzyl protection of commercially available 2-amino-3-
nitrophenol (16) and subsequent bromination of the resulting
intermediate 17 afforded 2-(benzyloxy)-4-bromo-6-nitroaniline
(18), which in turn was transformed with acetic anhydride into
the diacetyl derivate 19. The reduction of the nitro group and
concomitant cleavage of one acetyl group from substrate 19
was accomplished with a mixture of iron(III) chloride and
(1) The prochiral ketone 15 used as starting material for the
asymmetric hydrogenation reaction was prepared by transfor-
mation of the Mannich base 12 with 1-[1-(2-methylphenyl)vi-
nyl]pyrrolidine (13), which has a limited shelf life (Scheme 2).
Furthermore, the preparation of enamine 13 could not be
performed under standard conditions (acid, azeotropic removal
of water) and required the use of corrosive titanium tetrachloride.
Consequently, the possibility to accomplish the alkylation of
the Mannich base 12 using ethyl 3-(2-methylphenyl)-3-oxo-
propanoate (14) instead of the enamine 13 was investigated.
(2) The asymmetric reduction of ketone 15 was done with
a low S/C ratio of 100:1. Due to the high price of the catalyst
9, a cost-effective process could not be achieved under these
reaction conditions.
(3) Though there are some examples where the Mitsunobu
reaction has been applied successfully in process chemistry, the
removal of byproduct and the purification of the desired product
often turned out to be tedious.16
Here, we would like to report our results with respect to the
optimization of these reaction steps and the large-scale synthesis
of BYK 405879 (8).17 With exception of the Mitsunobu
cyclization, all reactions were performed in multikilogram scale.
An optimized procedure for the preparation of the building block
11 will be reported in due course.
(16) (a) Nishino, Y.; Komurasaki, T.; Yuasa, T.; Kakinuma, M.; Izumi,
K.; Kobayashi, M.; Fujiie, S.; Gotoh, T.; Masui, Y.; Hajima, M.;
Takahira, M.; Okuyama, A.; Kataoka, T. Org. Process Res. DeV. 2003,
7, 649–654. (b) Wallace, M. D.; McGuire, M. A.; Yu, M. S.;
Goldfinger, L.; Liu, L.; Dai, W.; Shilcrat, S. Org. Process. Res. DeV.
2004, 8, 738–743. (c) Dugger, R. W.; Ragan, J. A.; Brown Ripin,
D. H. Org. Process Res. DeV. 2005, 9, 253–258.
(17) Palmer, A.; Webel, M.; Scheufler, C.; Mu¨ller, B.; Haag, D.; Brehm,
C.; Chiesa, M. V.; Zimmermann, P. J.; Buhr, W. Patent Application
WO 2008/074858, 2008.
(18) Chiesa, V.; Palmer, A.; Brehm, C.; Grundler, G.; Senn-Bilfinger, J.;
Simon, W.-A.; Postius, S.; Kromer, W.; Buhr, W.; Zimmermann, P. J.
Patent Application WO 2004/054984, 2004.
1172
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

Scheme 4a
hydrazine hydrate and therefore was considered critical. How-
ever, the reaction was feasible on a multi kilogram scale and
the aniline derivative 20 was isolated in 60-80% yield. The
bis-methylation of 20 was performed by reductive amination
using formaldehyde and sodium cyanoborohydride. Although
the transformation resulted in a 70-90% yield, the concerns
related to the toxicity of sodium cyanoborohydride still need
to be addressed. Benzimidazole 22 was obtained by treatment
of diamine 21 with phosphorous oxychloride. Finally, the
synthesis of key intermediate 11 was completed by palladium-
catalyzed amidocarbonylation of bromide 22 under conditions
described in literature and hydrogenolytic cleavage of the benzyl
protecting group.19 In particular, the carbonylation step was
considered a major drawback due to the toxicity of carbon
monoxide and the fact that only few contract manufacturers
offer the special equipment necessary for handling carbon
monoxide on a reasonable scale. Furthermore, hydrogenolytic
debenzylation was sluggish and required an unacceptably high
amount of catalyst (up to 15%).
2.2. Large-Scale Synthesis of Ketone 15. For the prepara-
tion of ketone 15 Via the two routes proposed in Scheme 2, the
Mannich base 12 was required as starting material. Building
block 12 was prepared by the reaction of benzimidazole 11 with
Eschenmoser’s salt and successful transformation of 11 was
accomplished with both dimethylmethylideneammonium iodide
and dimethylmethylideneammonium chloride, leading to 12a
or 12b respectively.
The original protocol was not suitable for larger scale
because it used dichloromethane as a solvent and included a
laborious extractive workup procedure. Consequently, an op-
timized procedure was developed which employed a mixture
of water and 2-propanol as a solvent (Scheme 4). The addition
of triethylamine proved to be beneficial, as the reaction had
sometimes stalled and had not gone to completion. This protocol
allowed for the isolation of the precipitated product 12 directly
from the reaction mixture Via filtration or centrifugation and
its direct use for the next reaction step without any purification.
Despite these improvements, it quickly turned out that the
major drawback of this procedure was the insufficient supply
and the high price of Eschenmoser’s salt. Although several
suppliers claimed to be able to deliver kilogram quantities, we
often ran short of this raw material due to delivery problems.
Therefore, we examined the in situ generation of Eschenmoser’s
salt from formaldehyde and dimethylammonium chloride. This
procedure worked smoothly in pilot plant scale, using 2-pro-
panol as the solvent and triethylamine as the base and, after
the subsequent addition of 11, the expected product 12 was
obtained in comparable yields as when Eschenmoser’s salt was
employed.
a Reagents and conditions: i) Eschenmoser’s salt (1.3 equiv), Et3N (0.25
equiv), 2-PrOH, rt, 16 h, 100% 12a; (ii) [NMe2H2]+Cl- (1.3 equiv), Et3N (0.3
equiv), 2-PrOH, addition of aq formaldehyde (1.3 equiv), 35-45 °C, 1-3 h,
100% 12b; (iii) 14 (1.4 equiv), NaOH (2.0 equiv), toluene, H2O, rfl., 8 h; (iv)
citric acid, n-BuOH, H2O, Σ53%; (v) aq NH3 (25 wt %), CH2Cl2, H2O, 82%;
(vi) Na-tert-pentylate (25% in toluene, 2.3 equiv), toluene/DMF, 55-85 °C,
2.5-4 h, addition of water, reflux, 2-3 h, 90%.
acid with ethyl acetate or purchased from an external supplier,
showed that the alkylation of Mannich base 12 worked in
principle, but was hampered by disappointing yields of about
30% (Scheme 4). However, since the whole concept looked
promising, it was decided to put more effort into optimizing
the reaction. It was quickly recognized that one key factor,
particularly in regard to the following reduction step, was the
complete decarboxylation of intermediate 24 to ketone 15. As
specified further in section 2.3, erosion of the enantiomeric
excess was observed when starting material 15 containing
residual amounts of the nondecarboxylated intermediate 24 was
submitted to the asymmetric reduction step. Therefore, we had
to ensure reaction conditions leading to fully decarboxylated
product 15 which in turn seemed to compromise the yield. A
further problem arising in the course of the optimization was
the formation of the undesired byproduct 25. Ether 25 seemed
to arise from intermediate 24, involving an unexplained reduc-
tion of the carboxylic acid moiety. In later development phases,
For initial and fast supplies of API at the outset of the
development, enamine 13 was used for the preparation of ketone
15 (Scheme 2), but the handling disadvantages (oily liquid with
limited shelf life, difficult preparation) rendered this route not
particularly amenable for large scale; consequently, an alterna-
tive was sought. Brief examination of the use of ꢀ-ketoester
14, which was prepared by condensation of 2-methylbenzoic
(19) (a) Head, R. A.; Ibbotson, A. Tetrahedron Lett. 1984, 25, 5939–5942.
(b) Horino, H.; Sakaba, H.; Arai, M. Synthesis 1989, 715–718.
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1173

Scheme 5a
employing Mannich base 12b derived from in situ generated
Eschenmoser’s salt, byproduct 25 was no longer observed.
As a consequence, much optimization work was performed,
which included the choice of an appropriate base and solvent
and modifications of the reaction temperature and concentration
as well as the equivalents of the employed ketoester 14. Finally,
a preliminary procedure was found, which produced the desired
ketone 15 on kilogram scale. In this optimized procedure the
transformation of Mannich base 12 was performed in a biphasic
reaction mixture consisting of toluene/aqueous sodium hydrox-
ide solution, and the ketone 15 typically was obtained in
40-50% yields after precipitation as the citric acid salt,
subsequent release by treatment with a base, and crystallization
from acetone.
Although this process allowed the fast production of several
kilograms of 15 in our kilolabs, it was far from optimal. The
yields were not satisfying, the purification procedure was
laborious, and dichloromethane was used for extraction. There-
fore, a further stage of development was required and finally
achieved by the use of potassium tert-pentylate as a base in a
homogeneous mixture of toluene and DMF. In our pilot plants,
this procedure was easily carried out on a multikilogram scale
and gave a tremendous improvement in yield to over 90%.
Moreover, ketone 15 was obtained in excellent purity, and the
additional purification step with citric acid was no longer
required.
2.3. Optimization of the Asymmetric Hydrogenation
Reaction. The results obtained in the course of our previous
investigations indicated that the outcome of the asymmetric
hydrogenation depended on the substitution pattern of the
respective ketone 5 and on the amount of base employed.10 It
was also demonstrated that, in the absence of catalyst 9, the
base-catalyzed side reaction leading to products 10 was slow
at room temperature but fast at 70 °C.
a Reagents and conditions: (i) benzyl bromide (1.0-1.1 equiv), K2CO3
(1.0-1.1 equiv), DMF, rt or 65 °C, 4-17 h, 28: 83%, 29: 81%, 30: 90%; (ii)
RuCl2[(S)-Xyl-P-Phos][(S)-DAIPEN], t-BuOK; 2-PrOH, t-BuOH, H2, see Tables
1, 3 and 4; (iii) Method A: Pd/C, H2, EtOH or MeOH, rt, 3-18 h, 34: 89-95%,
35: 69%, 36: 82%; Method B: Pd/C, 1,4-cyclohexadiene, rt-55 °C, 4 h, 34:
92%.
(1) At a S/C ratio of 100:1, the reduction of ketone 15
proceeded in a reliable manner (entries 4-7) affording alcohol
34 in optical purities of 84.2-93.3% ee. Whereas the increase
of substrate concentration exerted a beneficial influence on the
enantioselectivity of the reduction (entry 6), the optical purity
of alcohol 34 was compromised in the presence of higher
amounts of potassium tert-butylate (entry 7). At a S/C-ratio of
250:1, only 66% conversion was achieved under standard
conditions (entry 13). Quantitative transformation of ketone 15
to alcohol 34 was achieved either by increasing the amount of
potassium tert-butylate (entry 14), the reaction time (entries 15,
16), the substrate concentration (entries 16, 17), and/or the
hydrogen pressure (entry 17). The optical purity of diol 34
depended on the substrate concentration of ketone 15. If 0.21
and 0.5 M solutions of ketone 15 were used for the hydrogena-
tion reaction, the diol 34 was isolated in an optical purity of
∼80% ee (entries 14 and 15) and ∼90% ee (entries 16 and
17), respectively. At a S/C ratio of 200:1, the reduction of ketone
15 proceeded in a reliable manner (225 g scale, entry 12)
affording alcohol 34 in 80% yield, 96.0% chemical purity, and
94.2% optical purity. According to the hydrogen uptake curve,
complete transformation of ketone 15 occurred within a period
of 8 h.
A total of 52 experiments was conducted to judge the
influence of pressure, reaction time, solvent, S/C ratio, substrate
concentration, amount of base, and the quality of the starting
material on the outcome of the reduction of the ketone 15 to
the alcohol 34 (Scheme 5). A representative summary of these
experiments is shown in Table 1.
An important finding was that the outcome of the reaction
depended on the quality of the starting material (Table 1, entries
1-3). The optical purity of the chiral alcohol 34 differed
significantly when two batches of ketone 15 possessing HPLC
purities of 98.9% and 95.3% were reduced under identical
reaction conditions. An ee value of 89.0% was determined for
the first batch of product 34 (entry 1), whereas the second batch
possessed an optical purity of only 61.7% ee (entry 2). Since
the HPLC chromatogram suggested the presence of intermediate
24 in the less pure batch, we investigated whether this derivate
was responsible for the unexpected drop in the enantioselec-
tivity. Indeed, when a mixture of ketone 15 (95%) and the
intermediate 24 (5%) was reduced in the presence of the catalyst
9 (S/C ) 100:1), the alcohol 34 was obtained with an optical
purity of only 59.9% ee (entry 3).
(2) As expected, the base-catalyzed background reaction was
favoured in the presence of higher amounts of base and at high
S/C ratios. At a S/C ratio of 100:1, pure batches of alcohol 34
were isolated (see Table 2). On the other hand, if the
The following conclusions can be drawn from the results
listed in Table 1:
1174
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

Table 1. Asymmetric hydrogenation of ketone 15 (RuCl2[(S)-Xyl-P-Phos][(S)-DAIPEN], 2-PrOH, t-BuOH, 65-70 °C)
entry scale (g)/additive purity of 15 (%)a S/C ratio t-BuOKb (equiv) concn (mol/L) p H2 (bar) time (h) convc (%) eed of 34 (%)
1
2
3
4
5
6
7
8
9
5
98.9
95.3
98.8
98.7
98.7
97.7
98.8
98.4
96.0
96.0
96.0
96.0
98.8
98.6
98.6
98.4
96.0
96.0
96.0
100:1
100:1
100:1
100:1
100:1
100:1
100:1
100:1
100:1
100:1
100:1
200:1
250:1
250:1
250:1
250:1
250:1
250:1
300:1
1.1
1.1
1.1
1.1
1.1
1.1
1.7
1.1
1.1
1.1
1.1
1.1
1.1
1.4
1.1
1.1
1.1
1.1
1.1
0.21
0.21
0.21
0.21
0.21
0.5
80
80
80
80
80
80
80
80
30
10
20
20
20
20
48
20
20
20
20
20
20
20
20
20
72
48
20
20
20
100
100
100
100
100
100
100
47
100
100
96
100
66
100
98
100
100
100
100
89.0
61.7
59.9
90.6
90.3
93.3
84.2
88.2
95.1
97.0
94.8
94.2
83.9
78.1
80.5
90.5
91.2
86.0
84.3
4
4/+5% 24
70
31
10
4.7
0.25
0.21
0.5
4/+10% H2O
10
10 10
11 10
12 225
13
14
15
16 10
17 10
18 17/without 2-PrOH
19 17/without 2-PrOH
0.5
0.5
5-8
0.5
80
80
4
4
4
0.21
0.21
0.21
0.5
80
80
80
0.5
120
80
0.9
0.9
120
a Purity determined by HPLC on the isolated product. b A 1 M solution of potassium tert-butylate in tert-butanol was used. c 87-98% crude product after workup
(distribution between CH2Cl2/saturated NH4Cl solution), conversion determined by HPLC. d Optical purity determined by capillary electrophoresis.
Table 2. Purification of crude alcohol 34 by crystallization
entry from ee method of yield chemical ee
entry Table 1 (crude; %) purificationa (%) purity (%) (%)
byproduct 10 which could be removed after Mitsunobu cy-
clization.
(3) Interestingly, lower enantioselectivities were obtained if
tert-butanol was used as the reaction solvent instead of a mixture
of 2-propanol and tert-butanol (compare entries 16 and 18). In
contrast to our findings from the imidazo[1,2-a]pyridine series,
the presence of water had a negative influence on the reaction
rate (entry 8).11
(4) The asymmetric reduction was also feasible applying low
hydrogen pressures of 5-8, 10, and 30 bar instead of 80 bar
(entries 9-11). On the other hand, the increase of hydrogen
pressure was beneficial for the outcome of the reaction at higher
S/C ratios (entries 17 and 19).
(5) High concentrations of starting material 15 were found
to be beneficial for a fast, clean, and enantioselective conversion
of ketone 15 to alcohol 34 (compare entries 4 and 6/entries 15
and 16). Using a high substrate concentration of 0.9 M, the
reaction was also feasible at a S/C ratio of 300:1 (entry 19).
Typically, the asymmetric reduction of ketone 15 afforded
alcohol 34 in an optical purity of 85-90% ee. As can be seen
from Table 2, several crystallization procedures for the purifica-
tion of crude 34 were developed.
The chemical purity was increased by crystallization of the
crude product from acetone (Table 2, entries 2-5). Alterna-
tively, alcohol 34 was converted into its oxalate salt and isolated
in pure form by subsequent treatment of the corresponding salt
with a base (Table 2, entry 1).
1
2
3
4
5
6
7
8
9
1
5
6
9
12
4
14
15
16
7
89.0
90.3
93.3
95.1
94.2
90.6
78.1
80.5
90.5
84.2
A
B
B
B
B
C
C
C
C
D
78
83
71
79
80
70
46
59
51
39
98.9
98.3
98.9
98.3
96.5b
99.4
94.0b
95.5b
95.5b
97.3
89.4
90.3
93.2
96.8
94.0
95.4
93.0
95.5
94.9
93.9
10
a Method A: crystallization from 2-propanol in the presence of oxalic acid (1.5
equiv) and treatment of the corresponding salt with CH2Cl2/NaHCO3 solution.
Method B: crystallization from acetone. Method C: crystallization from acetone in
the presence of L-(+)-mandelic acid (1.2 equiv) and treatment of the corresponding
salt with CH2Cl2/NaHCO3 solution. Method D: crystallization from ethyl methyl
ketone and further purification of the mother liquor according to method A.
b Samples contained 2.0-4.9% of byproduct 10.
Table 3. Asymmetric hydrogenation of O-protected ketones
28-30 (RuCl2[(S)-Xyl-P-Phos][(S)-DAIPEN], 2-PrOH,
t-BuOH)
S/C
t-BuOK
p
T
time concn yield
ee
entry ketone ratio
(equiv) (bar) (°C) (h) (mol/L) (%)
(%)a
1
2
3
4
5
6
7
8
28
28
28
28
28
28
28
28
28
28
28
28
28
29
29
30
100:1
100:1
250:1
350:1
500:1
1.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.4
100
100
100
80
80
80
80
80
80
80
80
10
10
80
80
80
70
70
70
65
65
70
70
70
70
70
70
70
70
70
70
70
17
17
17
20
20
18
18
18
18
18
18
18
18
17
17
17
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.29
0.29
0.11
68b
72b
59b
75c
72c
75c
95d
70c
-/99.5
-/98.8
-/98.8
98.5/100
98.7/99.2
97.7/99.7
98.6/-
700:1
1000:1
1500:1
2500:1
3500:1
5000:1
2000:1
3500:1
100:1
98.2/98.9
9
100d 98.2/-
10
11
12
13
14
15
16
66c
70c
73c
74c
82c
83c
49b
96.3/98.7
96.6/99.2
96.7/98.2
96.3/99.2
-/99.5
-/99.7
-/98.2
Two methods were available for the enhancement of both
the optical and chemical purity of alcohol 34. Treatment of the
crude product with L-(+)-mandelic acid afforded the corre-
sponding salt from which the pure alcohol 34 was released by
treatment with a base (Table 2, entries 6-9). Alternatively, an
increase of the optical purity was secured by crystallization of
crude 34 from ethyl methyl ketone. Interestingly, in this solvent,
the racemic mixture turned out to be less soluble than the (3R)-
enantiomer, and the alcohol 34 was obtained in 39% yield and
2000:1
50:1
a Optical purity reported for crude alcohol/purified sample. b Purification by
column chromatography. c Purification by crystallization from acetone. d Crude
product.
hydrogenation was conducted with S/C ratios of 200:1 or
250:1, the isolated samples of alcohol 34 contained 2-5% of
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1175

Table 4. Asymmetric hydrogenation of ketone 28 in a 10 L Premex Hastelloy autoclave (RuCl2[(S)-Xyl-P-Phos][(S)-DAIPEN] 9,
2-PrOH, t-BuOH, c ) 0.50 M, 80-100 bar H2, 70 °C, 20 h)
entry
1
ketone
t-BuOK (1 M solution)
catalyst
yield
1000 g/2.13mol
213 mL/0.21 mol
2.65 g (S/C ) 1000:1)
946 g/94%
2
3
1000 g/2.13mol
1000 g/2.13mol
213 mL/0.21 mol
213 mL/0.21 mol
1.06 g (S/C ) 2500:1)
1.06 g (S/C ) 2500:1)
1711 g/85%
2543 g/90%
4
5
6
1000 g/2.13mol
1000 g/2.13mol
818 g/1.74mol
213 mL/0.21 mol
213 mL/0.21 mol
174 mL/0.17 mol
1.06 g (S/C ) 2500:1)
1.06 g (S/C ) 2500:1)
0.90 g (S/C ) 2500:1)
93.9% ee by subsequent crystallization from the mother liquor
(Table 2, entry 10).
It was also shown that the high pressure of 80 bar that
typically had been used to increase the reaction rate was not
required (entries 12 and 13). At a hydrogen pressure of 10 bar
and a S/C ratio of 2000:1 and 3500:1, respectively, complete
conversion of ketone 28 was accomplished, and the alcohol 31
was isolated in 73-74% yield and 98.2-99.2% ee. However,
at a S/C ratio of 5000:1 the reduction did not go to completion
using the standard reaction time of 17 h and a lower hydrogen
pressure of 10 bar (35% conversion, 98.5% ee). These findings
are particularly important because they allow the use of standard
pilot-plant vessels for hydrogenation instead of the often size-
limited and specialized autoclave equipment, leading to a higher
throughput and a better cost-effectiveness.
Since the ketone 28 showed significantly higher reactivity
than its O-unprotected analogue 15, we decided to reinvestigate
the asymmetric hydrogenation of substrates that had previously
turned out to be challenging. In the case of ketones 26 and
27, the reaction rate of the hydrogenation was low, and
consequently, large quantities of byproduct 10 (Ar ) 2-meth-
ylphenyl) were formed.10 In contrast to these findings, the
asymmetric reduction of the benzyl-protected analogues 29 and
30 proceeded smoothly and afforded the corresponding enan-
tiopure alcohols 32 and 33 in good yields (Table 3, entries
14-16). In the case of ketone 29 containing an azetidine
carboxamide moiety, it was demonstrated that identical results
were obtained using S/C ratios of 100:1 and 2000:1, respectively
(entries 14 and 15). The reduction of ketone 30 possessing a
Weinreb amide substituent was only demonstrated at a S/C ratio
of 50:1 (entry 16).
Next, we repeated the asymmetric hydrogenation reaction
of ketone 28 in a 2 L Premex Hastelloy autoclave using a
substrate concentration of 0.18 M (32 g of starting material 28)
and a S/C ratio of 500:1 under otherwise identical reaction
conditions (0.1 equiv of potassium tert-butylate, 2-propanol/
tert-butanol, 80 bar hydrogen pressure, 70 °C, 20 h). The optical
purity of the crude product (97.4% ee) was enhanced by
crystallization from acetone, and the pure alcohol 31 was
isolated in 79% yield and 99.5% ee.
We then turned our attention to the asymmetric reduction
of ketone 28 on a kilogram scale using a 10 L Premex Hastelloy
autoclave, the largest autoclave available at our research site.
Since this autoclave was employed for different catalytic
reactions using a variety of transition metal complexes, we were
concerned about the presence of impurities that might reduce
the activity of the hydrogenation catalyst 9 or compromise the
enantioselectivity of the reduction. For this reason, we first
conditioned the reaction vessel by heating a solution of catalyst
9 in 2-propanol and then confirmed the successful removal of
potential impurities by asymmetric reduction of acetophenone
At this stage of the optimization of the asymmetric hydro-
genation reaction, the reduction of 1 kg of ketone 15 still
required 33 g of catalyst 9 (equivalent to a S/C ratio of 200:1).
By applying these conditions on a 225 g scale, the corresponding
alcohol 34 was isolated in 80% yield and 94% ee. Due to the
high price of the hydrogenation catalyst (220 Euros/g), this
option was economically not viable, and alternatives were
evaluated.
The low catalytic activity could either be attributed to the
chelating properties of the nitrogen atom of the benzimidazole
ring and the hydroxy group or to the steric bulk of the ortho-
methylphenyl substituent. In the first case, the introduction of
a base-resistant protecting group, preferably also stable towards
hydrolytic cleavage, was expected to diminish the chelating
properties of the substrate and promote a successful outcome
of the reaction in the presence of less hydrogenation catalyst.
Furthermore, protection of the acidic phenol moiety should
allow the use of substoichiometric amounts of base since
deprotonation of the phenol group prior to activation of the
catalyst no longer occurs.
Consequently, we decided to investigate the asymmetric
reduction of the benzyl-protected ketone 28 in the presence of
RuCl2[(S)-Xyl-P-Phos][(S)-DAIPEN] 9. The starting material
28 and the structurally related substrates 29 and 30 were
obtained by conversion of the respective phenolic ketone 15,
26 and 27 with benzyl bromide under basic conditions (potas-
sium carbonate) and reslurrying of the obtained crude products
(Scheme 5).
In order to determine the maximum S/C ratio for the
asymmetric reduction of the benzyl-protected ketone 28 in the
presence of catalyst 9, a series of hydrogenation experiments
was conducted in a 100 mL Premex Hastelloy autoclave using
10 g of starting material (Table 3). At a S/C ratio of 100:1,
almost identical results were obtained in the presence of
stoichiometric (entry 1) or catalytic amounts of potassium tert-
butylate (entry 2). The experiments listed in entry 2-11 were
performed in the presence of 0.1 equivalents of potassium tert-
butylate using varying amounts of catalyst 9. The enantiose-
lectivity of the reduction was determined at the stage of the
crude alcohol 31 and after purification, which was accomplished
either by column chromatography or by crystallization from
acetone. To our delight, up to a S/C ratio of 5000:1, neither the
yield nor the optical purity of alcohol 31 was compromised.
At a S/C ratio of 7500:1, the reaction rate decreased, and in
the course of 18 h, only 44% conversion to alcohol 31 took
place (80 bar hydrogen pressure, 70 °C). Also, the enantiose-
lectivity of the reduction was slightly reduced (94.4% ee).
1176
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

Scheme 6a
in the presence of catalyst 9 and catalytic amounts of potassium
tert-butylate. Since the corresponding hydrogenation product,
(1R)-1-phenylethanol, was obtained in excellent optical purity,
we examined the asymmetric reduction of ketone 28. As can
be seen from Table 4, clean conversion took place at S/C ratios
of 1000:1 and 2500:1, and the alcohol 31 was obtained in
85-94% yield. In the course of these experiments the workup
procedure was optimized: On a small scale, the reaction mixture
was quenched with saturated ammonium chloride solution, and
alcohol 31 was extracted with dichloromethane. At larger scale,
the reaction mixture was neutralized with acetic acid and diluted
with water at a temperature of 50-55 °C. The product 31
crystallized out and was isolated in a convenient manner. The
combined batches listed in Table 4 were crystallized from
2-propanol, and the alcohol 31 was isolated in 81% yield and
an optical purity of 98.5% ee.
a Reagents and conditions: (i) Method A (process development): PPh3 (1.3
equiv), DIAD (1.3 equiv), toluene, <25 °C, then salt formation with succinic
acid, 8: 80%; Method B (medicinal chemistry): PPh3 (1.3 equiv), DIAD (1.3
equiv), THF, rt, 0.5-1 h, 37: 59%, 99.2% ee; 38: 47%, 97.0% ee.
addition of aqueous ammonia solution, and the product 8
extracted with methylisobutyl ketone. Further purification was
accomplished by precipitation of the product as the succinic
acid salt from methylisobutyl ketone. This salt of target
compound 8 was isolated typically in over 80% yield and with
practically no deterioration in the enantiomeric excess overall
(as compared to the starting diol 34). Although minor erosion
of the optical purity seemed to take place during the cyclization
reaction, this finding is of no practical relevance since the
enantiomeric excess of the material could easily be upgraded
by subsequent salt formation and crystallization. The free base
was returned from the acid salt by treatment with aqueous
ammonia solution, extraction, and final crystallization from
isopropyl acetate, furnishing BYK 405879 (8) in excellent
purity.
Since the required amount of BYK 405879 (8) was supplied
on time, some minor issues associated with the Mitsunobu
cyclization that might be disadvantageous for further upscale
have not yet been eliminated, nor has the process been
transferred from our kilolabs to the pilot plant. For the same
reason, diol 34 with high enantiomeric excess derived from the
asymmetric reduction/deprotection of O-benzyl protected ketone
28 was not employed on a larger scale for the Mitsunobu
reaction (see Experimental Section). Furthermore, the prelimi-
nary results indicating that the asymmetric reduction of ketone
28 can be conducted under low pressure, and thus circumvent
the use of autoclave equipment, remain to be confirmed on a
larger scale.
Figure 3 illustrates the results of the X-ray analysis obtained
for the 2:1 salt of BYK 405879 (8) with succinic acid.
Interestingly, there were two possible structure solutions: a
centrosymmetric solution which would indicate a racemic
compound and a solution with two independent molecules of
8 and one molecule of succinic acid in the crystallographic
asymmetric unit. Although the structure showed a so-called
pseudosymmetric behaviour, in reality, the structure is indeed
chiral. In solid state, the system shows no ionic properties, and
the hydrogens of the dicarboxylic acid are clearly localized at
the oxygens of the acid functions. The hydroxy functions are
donors of a strong intermolecular hydrogen bond interaction
to the nitrogens of the imidazole moieties. The succinic acid
constitutes a bridge-like link between two molecules of BYK
405879 (8), forming nearly linear hydrogen bonds. The
Cleavage of the protecting group was then accomplished
either by catalytic hydrogenation (palladium on charcoal,
hydrogen) or by catalytic transfer hydrogenation (palladium on
charcoal, 1,4-cyclohexadiene). The hydrogenolysis proceeded
equally well with crude batches or pure batches of benzyl ether
31 (obtained by crystallization from acetone) and afforded the
diol 34 in 89-95% yield. The ruthenium and palladium content
of the 5 kg batch of diol 34 was analyzed, and values of 11
ppm (Ru) and 27 ppm (Pd) were determined. The heavy metal
content was reduced further in the course of the Mitsunobu
cyclization to BYK 405879 (8) and the subsequent purification
of the API, and always met the specification threshold for total
heavy metals of <20 ppm. In the same manner, hydrogenolytic
cleavage of the benzyl protecting group present in substrates
32 and 33 furnished the respective diols 35 and 36 in 69-82%
yields.
In summary, the catalyst loading could be reduced signifi-
cantly by the presence of the benzyl protecting group, and only
1 g of hydrogenation catalyst was required for the asymmetric
reduction of 1 kg of ketone 28. Both the introduction and the
cleavage of the protecting group were accomplished in good
yield applying standard reaction conditions.
2.4. Optimization of the Mitsunobu Cyclization. For the
ring-closure reaction of diol 34 to our development candidate
8, the originally used Mitsunobu conditions (DIAD/triph-
enylphosphine) initially seemed to be the best choice. However,
as mentioned before, the removal of the typical Mitsunobu
byproduct (diisopropyl hydrazine-1,2-dicarboxylate/triphe-
nylphosphine oxide) was troublesome and often required
chromatography; thus, a more convenient protocol had to be
established. A first look at the reaction itself, yielding normally
about 50-60% (cf. Scheme 6, target compounds 37 and 38
prepared in medicinal chemistry), revealed some potential for
improvement. It quickly turned out that yields could be
improved considerably by using toluene instead of THF as a
solvent together with careful control of the temperature of the
exothermic reaction.
With respect to the workup, an aqueous extraction with dilute
hydrochloric acid seemed to be most suitable to partly remove
the byproduct of the Mitsunobu reaction. The aqueous phase
was subsequently extracted with methylisobutyl ketone in order
to separate further impurities, the pH readjusted to 10-12 by
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1177

Figure 3. X-ray analysis of the salt of BYK 405879 (8) with succinic acid: (a) molecule A with X-ray numbering system; (b) molecule
B and succinic acid with X-ray numbering system; (c) hydrogen bond situation.
hypothesis that a perpendicular orientation of the aryl moiety
to the heterocyclic scaffold is required to effect potent inhibition
of the H+/K+-ATPase was supported by the results of the X-ray
analysis: The flat benzimidazole moiety of both molecules of
8 have a nearly perpendicular orientation to the methylphenyl
moiety with angles between these planes of 83.2° (molecule
A) and 76.5° (molecule B), respectively.
4. Experimental Section
General. All chemicals were purchased from the major
chemical suppliers and used without any further purification.
The syntheses of the hydrogenation catalyst 9 and the ketones
26 and 27 were performed using the methods described
previously.9-11 The progress of the reaction was monitored on
Macherey-Nagel HPTLC plates Nano-SIL 20 UV254 (0.20 mm
layer, nano silica gel 60 with fluorescence indicator UV254) using
dichloromethane/methanol as a solvent system. Column chro-
matography was performed with Merck silica gel 60 (70-230
mesh ASTM) with the solvent mixtures specified in the
corresponding experiment. Spots were visualized by iodine
vapour or by irradiation with ultraviolet light (254 nm). Melting
points (mp) were taken in open capillaries on a Büchi B-540
melting point apparatus and are uncorrected. 1H NMR spectra
were recorded with a Bruker DRX 200 FT-NMR spectrometer
at a frequency of 200 MHz, a Bruker AV 300 FT-NMR
spectrometer at a frequency of 300 MHz, or a Bruker AV 400
FT-NMR spectrometer at a frequency of 400 MHz. 13C NMR
spectra were acquired with a Bruker AV 400 FT-NMR
spectrometer at a frequency of 100 MHz. DMSO-d6 was used
as a solvent. The chemical shifts were reported as parts per
million (δ ppm) with tetramethylsilane (TMS) or DMSO as an
internal standard (1H: δTMS ) 0.00 ppm; 13C δDMSO ) 39.50
ppm). High-resolution mass spectra were obtained on an Agilent
LC/MSD TOF instrument using electrospray ionisation (ESI
positive). Elemental analysis was performed on a Carlo Erba
1106 C, H, N analyzer. The optical purity of the target
compounds and of selected intermediates was determined by
capillary electrophoresis (CE) and/or high pressure liquid
chromatography (HPLC). The experimental conditions for the
separation of the enantiomers are given for each example in
the Experimental Section. The heavy metal content of diol 34
was determined by inductively coupled plasma optical emissions
spectroscopy (ICP-OES) using a Perkin-Elmer Optima 3000
3. Conclusion
In summary, we have developed an efficient asymmetric
synthesis that permitted the preparation of kilogram quantities
of the potassium-competitive acid blocker BYK 405879 (8).
Suitable strategies were identified to overcome the challenges
associated with the medicinal chemistry route. The Mannich
base 12 was prepared using cheap reagents (dimethylammonium
chloride/formaldehyde) instead of expensive Eschenmoser’s salt.
Ketone 15 was obtained in good yield and reproducible quality
by alkylation of Mannich base 12 with ꢀ-ketoester 14. In
contrast to enamine 13 used previously for the preparation of
ketone 15, ꢀ-ketoester 14 is readily available and can be stored
for an indefinite period. By introducing a benzyl protecting
group, the asymmetric ketone reduction could be performed at
high S/C ratios, affording alcohol 31 in high yields and excellent
optical purities. After cleavage of the protecting group, BYK
405879 (8) was obtained by Mitsunobu reaction and could be
purified in an effective manner by acid-base extraction and
crystallization in the presence of suitable acids. Finally, most
steps, except for the asymmetric reduction and the Mitsunobu
cyclization, were conducted successfully at a pilot-plant scale.
Despite the three additional steps required for the introduction
and removal of the benzyl protecting group and the final
crystallization of the API 8 in the presence of succinic acid,
the overall yield was improved considerably in comparison to
that of the medicinal chemistry route (43% versus 25% overall
yield based on benzimidazole 11).
1178
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

DV instrument with Perkin-Elmer ICP WinLab software.
Sample preparation was done by digestion of the sample with
nitric acid (65% m/v, 170 °C, 3 h). The content was calculated
by an external linear calibration line according to the monograph
2.2.22 “atomic emission spectrometry”, method 1 of the Ph.
Eur. fifth edition. Ruthenium and palladium standard solutions
(1000 mg/L) were used as a reference. The heavy metal content
of the final API 8 was determined according to the guidelines
(Ph. Eur. 2.4.8 method C). Karl Fischer titration was performed
on a Metrohm KF Coulometer 756 KF (Metrohm 774 Sample
Oven Processor).
(8S)-N,N,2,3-Tetramethyl-8-(2-methylphenyl)-3,6,7,8-tet-
rahydrochromeno[7,8-d]imidazole-5-carboxamide (8). To a
suspension of diol 34 (obtained by direct asymmetric hydro-
genation of ketone 15 in the presence of catalyst 9, 565 g, 1.48
mol, 91.6% ee) and triphenylphosphine (505 g, 1.93 mol) in
toluene (14 L), was added dropwise DIAD (380 mL, 1.93
mmol). The temperature was maintained below 25 °C in the
course of the addition, and an auburn solution was obtained.
After reaction completion, the mixture was extracted with
hydrochloric acid (1 N, 2 × 2.5 L). The aqueous phase was
washed with methylisobutyl ketone (3 × 1 L). The combined
organic phases were discarded. The pH value of the aqueous
phase was adjusted to 10-12 by addition of 25% aqueous
ammonia solution, and the product was extracted with meth-
ylisobutyl ketone (2 × 2 L). The combined organic phases were
concentrated under reduced pressure, yielding the crude title
compound (750 g).
Salt Formation with Succinic Acid. The crude product (750
g) was dissolved in methylisobutyl ketone (2.2 L). Succinic acid
(96 g, 0.815 mol) was added and the mixture heated to 80 °C
for several min. The suspension was allowed to cool to room
temperature, and stirring was continued for 17 h. The precipitate
was isolated by filtration and dried in Vacuo at 50 °C. This
afforded the title compound as a succinate salt (525 g of a
colorless solid, 83% yield, 90.2% ee, HPLC purity >97%,
stoichiometric ratio with respect to succinic acid, 1:0.53); mp
192-194 °C. 1H NMR (DMSO-d6, 400 MHz): δ ) 1.99 (mc,
1 H), 2.22 (mc, 1 H), 2.38 (s, 3 H), 2.42 (s, 4 H), 2.47 (s, 3 H),
2.65 (mc, 1 H), 2.81, 2.91 (s, mc, 4 H), 3.02 (s, 3 H), 3.68 (s,
3 H), 5.32 (dd, 1 H), 6.92 (s, 1 H), 7.24 (mc, 3 H), 7.47 (mc,
1 H).
Preparation of the Title Compound from Its Salt with
Succinic Acid. The salt of the title compound with succinic acid
(876 g, 2.05 mol) was suspended in dichloromethane (2.0 L),
and water was added (0.5 L). To this suspension, 25% aqueous
ammonia solution (0.4 L) was added under stirring. The mixture
was stirred for further 0.5 h, the organic phase separated, and
the aqueous phase extracted with dichloromethane (3 × 0.5 L).
The combined organic phases were washed with water (0.2 L),
coevaporated with methylisobutyl ketone (2 × 1.0 L), and the
residue was dissolved in isopropyl acetate (1.0 L) at 80 °C.
The solution was allowed to cool to room temperature over
3 h while the product precipitated. The precipitate was isolated
by filtration and dried in Vacuo at 50 °C. This afforded the title
compound (682 g of a colorless solid, 92% yield, 90.2% ee,
HPLC purity >99%); mp 179-181 °C. Determination of the
optical purity by HPLC (column: Daicel Chiralpak AD-H, 250
mm × 4.6 mm, 5 µm; eluant: n-heptane/ethanol, 85:15; flow
rate: 1 mL/min; detection wavelength: 218 nm): tR [(8R)-
enantiomer] ) 5.3 min/4.91 area %, tR [(8S)-enantiomer] )
1
6.7 min/95.09 area %, 90.2% ee. H NMR (DMSO-d6, 400
MHz): δ ) 1.98 (mc, 1 H, 7-Ha), 2.22 (mc, 1 H, 7-Hb), 2.38 (s,
3 H, C6H4-CH3), 2.49 (s, 3 H, 2-CH3), 2.62 (bs, 1 H, 6-Ha),
2.81 (s, 3 H, CON(CH3)2), 2.90 (bs, 1 H, 6-Hb), 3.02 (s, 3 H,
CON(CH3)2), 3.67 (s, 3 H, 3-CH3), 5.32 (dd, 3J7,8 ) 10.4 Hz,
3J7,8 ) 1.9 Hz, 1 H, 8-H), 6.92 (s, 1 H, 4-H), 7.24 (mc, 3 H,
C6H4-CH3), 7.47 (mc, 1 H, C6H4-CH3). 13C NMR (DMSO-
d6, 100 MHz): δ ) 13.2 (2-CH3), 18.6 (C6H4-CH3), 22.2 (C-
6), 27.8 (C-7), 29.7 (3-CH3), 33.9, 37.9 (CON(CH3)2), 74.3 (C-
8), 99.4 (C-4), 125.6, 125.9, 127.5, 130.2 (C6H4-CH3), 109.7,
131.0, 131.5, 134.8, 135.3, 139.3, 145.7, 151.2 (4° carbon
atoms), 170.0 (CON(CH3)2). Anal. Calcd for C22H25N3O2: C,
72.70; H, 6.93; N, 11.56. Found: C, 72.71; H, 6.90; N, 11.55.
HRMS (ESI) m/z C22H26N3O2 [M + H]+ Calcd: 364.2020.
Found: 364.2010. For the determination of heavy metals and
water content, an exemplary batch used for toxicity studies was
analyzed: heavy metals, total: <20 ppm; water content (Karl
Fischer titration): 0.02%.
5-[(Dimethylamino)methyl]-4-hydroxy-N,N,1,2-tetrame-
thyl-1H-benzimidazole-6-carboxamide hydroiodide (12a).
Benzimidazole 11 (1.5 kg, 6.43 mmol) was suspended in a
mixture of triethylamine (220 mL, 1.58 mol) and 2-propanol
(12.0 L). Eschenmoser’s salt (dimethylmethylidene ammonium
iodide; 1.55 kg, 8.36 mol) was added and the suspension stirred
for 16 h at room temperature. The precipitate was isolated by
filtration and dried in Vacuo (50 °C), yielding a mixture of the
title compound with 2-propanol (10 mol %) and triethylamine
(1 mol %): 2.8 kg, quantitative yield. This transformation was
also feasible using dimethylmethylidene ammonium chloride
1
as a reagent. H NMR (DMSO-d6, 400 MHz): δ ) 1.04 (d,
2-PrOH), 1.21 (t, NEt3), 2.59 (s, 3 H), 2.77 (s, 6 H), 2.96 (s, 3
H), 3.10 (s, 3 H), 3.77 (s, 3 H), 4.25 (s, 2 H), 7.17 (s, 1 H),
exchangeable protons not visible. 13C NMR (DMSO-d6, 100
MHz): δ ) 8.6 (NEt3), 13.3, 25.4 (2-PrOH), 30.2, 34.4, 35.0,
42.3, 45.8 (NEt3), 53.9, 61.9 (2-PrOH), 100.6, 106.3, 131.4,
137.0, 148.2, 152.7, 170.6. HRMS (ESI) m/z C15H23N4O2 [M
+ H]+ Calcd: 291.1816. Found: 291.1815.
5-[(Dimethylamino)methyl]-4-hydroxy-N,N,1,2-tetrame-
thyl-1H-benzimidazole-6-carboxamide hydrochloride (12b).
A mixture of benzimidazole 11 (5.0 kg, 21.43 mol) and
dimethylammonium chloride (2.3 kg, 28.20 mol) in triethyl-
amine (0.65 kg, 6.42 mol) and 2-propanol (35.0 L) was heated
to 35-45 °C, and formaldehyde (37% in water, 2.3 kg, 28.34
mol) was added over a period of 1-3 h at this temperature.
While adding the formaldehyde, the reaction mixture was
inoculated with several grams of product. After complete
addition of the formaldehyde, the mixture was stirred further
1-3 h at 35-45 °C. Methylisobutyl ketone (35.0 L) was added,
and 15 L of distillate was removed at 35-60 °C in Vacuo. The
reaction mixture was cooled to 10-20 °C and stirred for a
minimum of 1 h at this temperature. The precipitate was isolated
by filtration and dried in Vacuo (50 °C), yielding a mixture of
the title compound with isopropanol (1 mol %) and triethy-
1
lamine (10 mol %): 7.0 kg, quantitative yield. H NMR
(DMSO-d6, 400 MHz): δ ) 1.04 (d, 2-PrOH), 1.22 (t, NEt3),
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1179

2.58 (s, 3 H), 2.73 (s, 6 H), 2.93 (s, 3 H), 3.04 (q, NEt3), 3.08
(s, 3 H), 3.75 (s, 3 H, mc, 2-PrOH), 4.25 (s, 2 H), 7.13 (s, 1 H).
13C NMR (DMSO-d6, 100 MHz): δ ) 8.3 (NEt3), 13.2, 25.4
(2-PrOH), 30.0, 33.8, 34.9, 42.1, 45.1 (NEt3), 53.4, 100.5, 106.2,
131.5, 131.6, 137.0, 148.3, 152.6, 170.4. HRMS (ESI) m/z
C15H23N4O2 [M + H]+ Calcd: 291.1816. Found: 291.1824.
4-Hydroxy-N,N,1,2-tetramethyl-5-[3-(2-methylphenyl)-3-
oxopropyl]-1H-benzimidazole-6-carboxamide (15). Method
A. Mannich base 12b (800 g, 2.34 mol) was dissolved in 2 N
sodium hydroxide solution (2.3 L) and water (2.0 L) and added
over a period of 1 h to a refluxing solution of ethyl 3-(2-
methylphenyl)-3-oxopropanoate (669 g, 3.24 mol) in water (4.0
L) and toluene (4.0 L). The reaction mixture was stirred for
8 h at reflux. At room temperature, the precipitate was isolated
by filtration and dried in Vacuo (50 °C), yielding 902 g of the
crude title compound.
Salt Formation with Citric Acid. Over a period of 1 h, an
aqueous solution of citric acid (1.3 M, 4.4 L) was added to a
suspension of the crude product (902 g, 2.38 mol) in n-butanol
(5.0 L). The suspension was stirred for 60 h at room temper-
ature. The salt of the title compound with citric acid was isolated
by filtration and dried in Vacuo (50 °C): 778 g (1.26 mol) of a
colorless solid, 53% yield (stoichiometric ratio with respect to
citric acid: 1:1.75). 1H NMR (DMSO-d6, 400 MHz): δ ) 2.42
(s, 3 H), 2.52 (s, 3 H), 2.80 (s, 3 H), 3.00, 3.07 (s, bmc, 7 H),
3.67 (s, 3 H), 6.78 (s, 1 H), 7.30 (mc, 2 H), 7.42 (mc, 1 H),
7.70 (mc, 1 H), 10.00 (bs, 1 H).
Preparation of the Title Compound from Its Salt with Citric
Acid. A suspension of the salt of the title compound with citric
acid (1020 g) in dichloromethane (6.0 L) and water (4.0 L)
was neutralized with 25% aqueous ammonia solution (∼400
mL, pH 8-9). The phases were separated, and the aqueous
phase was extracted with dichloromethane (4.0 L). The com-
bined organic phases were washed with water (1.0 L, and the
solvent was evaporated. The residue was suspended in hot
acetone (1.2 L), the suspension was cooled down, and the
product was isolated by filtration and dried in Vacuo (50 °C).
The pure title compound was obtained in 82% yield (553 g,
HPLC purity >96%).
Method B. The inertized reaction vessel was charged with
Mannich base 12b (5.0 kg, 15.3 mol), ꢀ-ketoester 14 (3.8 kg,
18.4 mol), and 35 L of toluene. The suspension was heated to
55-62 °C, and a solution of potassium tert-pentylate (25% in
toluene, 17.4 kg, 34.5 mol) in DMF (10 L) was added over a
period of 2-3 h. The mixture was heated to 75-85 °C, stirred
at this temperature for 0.5-1 h, and diluted with water (40 L).
After a period of 2-3 h at reflux, 54 L of the solvents was
stripped off, the reaction mixture was cooled to 15-25 °C, and
stirring was continued for 2-5 h at this temperature. The
suspension was cooled further to 7 to -15 °C, stirred for another
1-2 h, and centrifuged. After drying in Vacuo at 50 °C, 5.2 kg
(90% yield) of the title compound was obtained; mp 209-211
°C. 1H NMR (DMSO-d6, 200 MHz): δ ) 2.42 (s, 3 H), 2.48
(s), 2.77, 2.80, 3.00, 3.05 (s, bmc, s, bmc, 10 H), 3.67 (s, 3 H),
6.78 (s, 1 H), 7.30 (mc, 2 H), 7.42 (mc, 1 H), 7.70 (mc, 1 H),
10.00 (bs, 1 H). 13C NMR (DMSO-d6, 100 MHz): δ ) 13.0,
20.6, 22.4, 29.8, 34.0, 38.4, 42.0, 98.0, 115.2, 125.8, 128.4,
131.1 (2×), 131.6, 132.0, 135.2, 136.8, 137.7, 145.9, 151.2,
170.7, 203.9. Anal. Calcd for C22H25N3O3: C, 69.64; H, 6.64;
N, 11.07. Found: C, 69.76; H, 6.62; N, 10.83. HRMS (ESI)
m/z C22H26N3O3 [M + H]+ Calcd: 380.1969. Found: 380.1964.
4-Benzyloxy-N,N,1,2-tetramethyl-5-[3-(2-methylphenyl)-
3-oxopropyl]-1H-benzimidazole-6-carboxamide (28). The
inertized reaction vessel was charged with ketone 15 (5.0 kg,
13.2 mol), potassium carbonate (2.0 kg, 14.5 mol), and DMF
(10 L). The suspension was heated to 50-58 °C, and benzyl
chloride (1.83 kg, 14.5 mol) was added over a period of 1.5-2
h. The mixture was stirred for 1.5-3 h at this temperature, and
aqueous ammonia solution (25%, 0.45 kg) was added over a
period of 10-30 min. Subsequently, cyclohexane (20 L) was
added within 15-30 min followed by addition of water (40 L)
over a period of 1-1.5 h. The suspension was stirred further
for 0.5-1 h, cooled to 8-15 °C over 3-24 h, and filtrated
over an agitated pressure filter dryer. The filtrate was discarded.
To the crude product on the filter dryer were added cyclohexane
(20 L) and water (40 L), and the suspension was stirred for
0.5-1 h at 30-40 °C and for another 30 min at 8-15 °C. The
title compound was isolated by filtration and dried in Vacuo
1
(50 °C): 5.3 kg (85% yield); mp 146-148 °C. H NMR
(DMSO-d6, 400 MHz): δ ) 2.36 (s, 3 H), 2.56 (s, mc, 4 H),
2.74 (s, 3 H), 2.95, 2.99 (bs, s, 6 H), 3.71 (s, 3 H), 5.80 (s, 2
H), 7.03 (s, 1 H), 7.27 (mc, 5 H), 7.40 (mc, 3 H), 7.52 (mc, 1
H). 13C NMR (DMSO-d6, 100 MHz): δ ) 13.4, 20.5, 22.6,
29.8, 34.0, 38.4, 42.3, 73.2, 101.2, 119.4, 125.7, 127.7, 127.8,
128.2 (2×), 131.0, 131.5 (2×), 133.2, 136.4, 136.6, 137.6,
137.9, 146.3, 151.6, 170.2, 203.6. Anal. Calcd for C29H31N3O3:
C, 74.18; H, 6.65; N, 8.95. Found: C, 73.53; H, 6.74; N, 8.82.
HRMS (ESI) m/z C29H32N3O3 [M + H]+ Calcd: 470.2438.
Found: 470.2445.
4-Benzyloxy-5-[(3R)-3-hydroxy-3-(2-methylphenyl)propyl]-
N,N,1,2-tetramethyl-1H-benzimidazole-6-carboxamide (31).
In a 10 L autoclave equipped with a glass inlay and filled with
nitrogen, ketone 28 was suspended in dry 2-propanol (entries
1-5: 4000 mL; entry 6: 3500 mL). Potassium tert-butylate
solution (1 M in tert-butanol) and the hydrogenation catalyst
RuCl2[(S)-Xyl-P-Phos][(S)-DAIPEN] were added (for quanti-
ties, see Table 4). The autoclave was purged with hydrogen
(3×), and the reaction mixture was hydrogenated at 70 °C and
80-100 bar pressure for 20 h.
Workup of Sample 1. The reaction mixture was cooled
to 35 °C, transferred into a glass vessel, neutralized with
acetic acid (12.2 mL, addition at a temperature of 65 °C),
and diluted with water (8 L, addition over a period of
0.25 h). The solution was cooled to 15-20 °C, and
precipitate was formed. The title compound was isolated
by filtration under pressure and dried in Vacuo at a
temperature of 50 °C (682 g of a colorless solid, 68%
yield). The filtrate was concentrated (removal of 2.5 L of
2-propanol/water). The warm solution was transferred into
a glass vessel and gradually cooled to 10 °C. Further 264 g
of the title compound (26% yield) was isolated by filtration
under pressure.
Workup of Samples 2 and 3. In the autoclave, the reaction
mixtures were neutralized by addition of acetic acid (12.2
mL each) and transferred into a 60 L vessel. At a
temperature of 50-55 °C, 24 L of water was added over
1180
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development

a period of 1 h. Stirring was continued for 1 h, and the
mixture was gradually cooled to 15-20 °C. The title
compound was isolated by filtration under pressure and
dried in Vacuo at a temperature of 50 °C (1711 g of a
colorless solid, 85% yield).
mol) was added a suspension of ketone 15 (225 g, 0.59
mol) in 2-propanol (520 mL). The mixture was stirred
for 90 min at room temperature and the hydrogenation
catalyst 9 (3.70 g, 3.0 mmol) added. After an additional
30 min, the dark-brown solution was transferred into a
2 L Premex Hastelloy autoclave with glass inlay. The
reactor was purged with hydrogen (3×) and subjected to
a hydrogen pressure of 80 bar. After a reaction time of
20 h at 70 °C, the autoclave was cooled and the reaction
mixture poured on a stirred mixture of saturated am-
monium chloride solution (2.2 L) and dichloromethane
(3.3 L). The phases were separated, and the aqueous phase
was extracted with dichloromethane (2 × 300 mL). The
combined organic phases were washed with water (2 × 1
L) and dried over sodium sulfate, and the solvent was
evaporated under reduced pressure. The crude title com-
pound (250 g of a green foamy solid, HPLC purity: 94.0%,
optical purity: 94.2% ee) was crystallized from acetone
(500 mL). The suspension was concentrated to a volume
of 400 mL, stirred for 17 h at room temperature, and
diluted with methyl tert-butyl ether (400 mL). After
additional 2 h at 0 °C, the title compound was isolated
by filtration, washed with methyl tert-butyl ether (200 mL)
and dried in Vacuo: 184 g of an off-white solid (80% yield,
HPLC purity: 96.5% containing 2.0% of byproduct 10,
optical purity: 94.0% ee); mp 205-207 °C. Determination
of the optical purity by CE (capillary: Agilent barefused
silica bubble, 56.0/64.5 cm × 50 µm; buffer: 50 mM
sodium phosphate (pH 2.5); chiral selector: 40 mM
heptakis(2,3,6-tri-O-methyl)-ꢀ-cyclodextrin; voltage: 30
kV; temperature: 20 °C; detection: diode array 219 nm):
tM [(3S)-enantiomer] ) 19.8 min; tM [(3R)-enantiomer]
Workup of Samples 4-6. Samples 4-6 were purified as
described above for samples 2 and 3.
The three batches listed in Table 4 were combined and
suspended in acetone (20 L). The suspension was stirred for
4.5 h at 50 °C, for 17 h at room temperature, and for 1.5 h at
10 °C. The title compound was isolated by filtration and dried
in Vacuo (50 °C): 4236 g of a colorless solid, 73% yield. The
mother liquor was concentrated to 25% of its original volume
and stirred for 17 h at room temperature. Another batch of the
title compound was obtained by filtration: 455 g of a colorless
solid, 8% yield; mp 160-163 °C. Determination of the optical
purity by HPLC (column: Daicel Chiralcel OD-H, 250 mm ×
4.6 mm, 5 µm; eluant: n-heptane/ethanol,: 90:10; flow rate: 1
mL/min; 40 °C; detection wavelength: 218 nm): tR [(3R)-
enantiomer] ) 18.2 min/98.14 area %, tR [(3S)-enantiomer] )
1
20.9 min/0.78 area %, 98.4% ee. H NMR (DMSO-d6, 400
MHz): δ ) 1.60 (mc, 1 H), 1.80 (mc, 1 H), 2.16 (s, 3 H), 2.54
(s, mc, 4 H), 2.67 (s, 3 H), 2.94 (s, mc, 4 H), 3.68 (s, 3 H), 4.67
(bs, 1 H), 5.00 (bs, 1 H), 5.71 (s, 2 H), 6.96 (s, 1 H), 7.11 (mc,
3 H), 7.34 (mc, 4 H), 7.44 (mc, 2 H). 13C NMR (DMSO-d6,
100 MHz): δ ) 13.4, 18.4, 24.3, 29.7, 33.9, 38.3, 38.7, 69.1,
73.2, 101.1, 120.9, 125.4, 125.5, 126.1, 127.5, 127.6, 128.1,
129.7, 131.6, 133.4, 133.6, 136.1, 138.1, 144.0, 146.3, 151.4,
170.3. HRMS (ESI) m/z C29H34N3O3 [M + H]+ Calcd:
472.2595. Found: 472.2592.
4-Hydroxy-5-[(3R)-3-hydroxy-3-(2-methylphenyl)propyl]-
N,N,1,2-tetramethyl-1H-benzimidazole-6-carboxamide (34).
Synthesis by Deprotection of Diol 31. The inertized reaction
vessel was charged with benzylated diol 31 (5.0 kg, 10.6
mol), palladium on charcoal (10%, type E101 Degussa,
50% H2O, 250 g), and 40 L of methanol. Under vigorous
stirring, the reactor was purged with hydrogen and the
reaction mixture hydrogenated at 15-30 °C and 3-4 bar
pressure for 3-5 h. Methanol (25 L) was added and the
mixture heated to 45-55 °C. Subsequently, the mixture
was filtrated through a layer of Hyflo Super Cel (1.5 kg)
in a heated pressure filter (45-55 °C). The filter cake was
washed with warm methanol (2 × 3 L), the complete
filtrate was transferred back to the reaction vessel, and
53-57 L of the solvent was distilled off in Vacuo at 50-60
°C. Crystallization was induced by inoculation of the
solution with 5-10 g of seed crystals of product 34. The
mixture was further stirred for 1-1.5 h at 50-60 °C
followed by addition of water (20 L). Another 5-7 L of
the solvents was stripped off in Vacuo at 50-60 °C. The
suspension was cooled to 8-15 °C within 1-3 h, stirred
for a minimum of 8 h, and finally centrifuged. The filter
cake was washed with water (5-8 L), furnishing 3.6 kg
(90% yield) of the title compound after drying in Vacuo
at 50 °C.
1
) 20.7 min; 94.0% ee. H NMR (DMSO-d6, 400 MHz):
δ ) 1.70 (bs, 1 H, 2′-Ha), 1.89 (bs, 1 H, 2′-Hb), 2.22 (s,
3 H, C6H4-CH3), 2.30-3.00 (m, 2 H, 1′-Ha, 1′-Hb), 2.51
(s, 3 H, 2-CH3), 2.70 (s, 3 H, N(CH3)2), 2.94 (bs, 3 H,
N(CH3)2), 3.65 (s, 3 H, 3-CH3), 4.71 (bs, 1 H, 3′-H), 5.04
(bs, 1 H, OH), 6.73 (s, 1 H, 4-H), 7.09 (mc, 2 H,
C6H4-CH3), 7.16 (mc, 1 H, C6H4-CH3), 7.42 (mc, 1 H,
C6H4-CH3), 10.18 (bs, 1 H, OH). 13C NMR (DMSO-d6,
100 MHz): δ ) 13.0 (2-CH3), 18.5 (C6H4-CH3), 24.0 (C-
1′), 29.7 (3-CH3), 33.9 (CON(CH3)2), 38.3 (CON(CH3)2,
C-2′), 69.1 (C-3′), 97.7 (C-4), 125.3, 125.5, 126.1, 129.7
(C6H4-CH3), 116.3, 131.2, 132.1, 133.7, 135.0, 144.2,
145.8, 150.9 (4° carbon atoms), 170.9 (CON(CH3)2). Anal.
Calcd for C22H27N3O3: C, 69.27; H, 7.13; N, 11.02. Found:
C, 68.98; H, 7.05; N, 10.52. HRMS (ESI) m/z C22H28N3O3
[M + H]+ Calcd: 382.2125. Found: 382.2123.
Acknowledgment
We are grateful to SynphaBase (Muttenz, Switzerland)
for providing us with kilogram quantities of benzimidazole
11, Ms. Tanja Ramisch (Nycomed, Konstanz) for the
determination of the optical purity of target compounds
and intermediates by capillary electrophoresis and HPLC,
Mr. Wolfgang Frey (University of Stuttgart) for the X-ray
analysis of the salt of BYK 405879 (8) with succinic acid,
Mr. Burkhard Grobbel, Mr. Bernd Marenski, Ms. Melanie
Horn and Mr. Oliver Meyer (Nycomed, Konstanz) for
Synthesis by Direct Catalytic Hydrogenation of Ketone 15.
To a flame-dried flask filled with argon and potassium
tert-butylate (1 M solution in tert-butanol, 652 mL, 0.65
Vol. 12, No. 6, 2008 / Organic Process Research & Development
•
1181

technical assistance, and to the Institute Fresenius (Berlin)
for analysing the ruthenium and palladium content of
selected samples.
12, 15, 28-38. 13C NMR spectra of compounds 8, 12,
15, 28, 29, 31, 32, 34, 35, 37, 38. Details of the X-ray
analysis of the salt of BYK 405879 (8) with succinic acid.
This information is available free of charge Via the Internet
at http://pubs.acs.org.
Supporting Information Available
Experimental procedures for the synthesis of com-
pounds 11, 17-23, 29, 30, 31 (screen of reaction
conditions and asymmetric hydrogenation in 32 g scale),
32, 33, 34 (methods of purification), 35-38. Analytical
Received for review July 29, 2008.
OP800177X
1
data of byproduct 25. H NMR spectra of compounds 8,
1182
•
Vol. 12, No. 6, 2008 / Organic Process Research & Development