Highly enantioselective catalytic asymmetric synthesis of a (R)-sibutramin precursor

Article abstract of DOI:10.2533/chimia.2010.59

The first highly enantioselective, catalytic asymmetric synthesis of di-des-methylsibutramine 3 is described. Dienamide 10, prepared by acetic acid anhydride quenching of the condensation product of nitrile 4 with a methallyl magnesium chloride, proved to

Full text of DOI:10.2533/chimia.2010.59

From ChemiCal researCh to industrial appliCationsꢀ
CHIMIAꢀ2010,ꢀ64,ꢀNo.ꢀ1/2ꢀ 59
doi:10.2533/chimia.2010.59ꢀꢀ
Chimiaꢀ64ꢀ(2010)ꢀ59–64ꢀ ©ꢀSchweizerischeꢀChemischeꢀGesellschaft
Highly Enantioselective Catalytic
Asymmetric Synthesis of a (R)-Sibutramin
Precursor
UlrichꢀBerens*a,ꢀAndreasꢀHafner*b,ꢀOliverꢀDosenbachb,ꢀTanjaꢀTritschlerb,ꢀFranzꢀSchwarzenbachb,ꢀ
Hans-JörgꢀKirnerb,ꢀChristopheꢀMalanc,ꢀandꢀOanhꢀMai-Huynhc
DedicatedꢀtoꢀProf.ꢀDr.ꢀDanielꢀBellušꢀonꢀtheꢀoccasionꢀofꢀhisꢀ70thꢀbirthday
Abstract:ꢀTheꢀfirstꢀhighlyꢀenantioselective,ꢀcatalyticꢀasymmetricꢀsynthesisꢀofꢀdi-des-methylsibutramineꢀ3ꢀisꢀde-
scribed.ꢀDienamideꢀ10,ꢀpreparedꢀbyꢀaceticꢀacidꢀanhydrideꢀquenchingꢀofꢀtheꢀcondensationꢀproductꢀofꢀnitrileꢀ4ꢀ
withꢀaꢀmethallylꢀmagnesiumꢀchloride,ꢀprovedꢀtoꢀbeꢀanꢀexcellentꢀsubstrateꢀforꢀruthenium-catalyzedꢀasymmetricꢀ
hydrogenationꢀwithꢀatropisomericꢀdiphosphineꢀligands.ꢀHydrogenationꢀwithꢀaꢀruthenium/(R)-ꢀMeOBiPhePꢀcatalystꢀ
atꢀS/Cꢀ=ꢀ500,ꢀgaveꢀtheꢀchiralꢀamideꢀ(R)-9ꢀinꢀ98.5%ꢀee inꢀalmostꢀquantitativeꢀyield.ꢀAfterꢀacidicꢀamideꢀhydrolysisꢀ
theꢀdesiredꢀamineꢀ(R)-3ꢀwasꢀobtainedꢀwithoutꢀerosionꢀofꢀenantioselectivity.ꢀItꢀisꢀanticipatedꢀthatꢀtheꢀoverallꢀpro-
cessꢀwillꢀbeꢀamenableꢀtoꢀlarge-scaleꢀproduction.ꢀ
Keywords:ꢀCatalyticꢀasymmetricꢀhydrogenationꢀ·ꢀEnantioselectiveꢀsynthesisꢀ·ꢀSibutramine
Introduction
(R)-2 might be useful for the treatment of asymmetric synthesis of (R)-2 by the enan-
CNS disorders.[2] Also, the enantiomers tioselective addition of isobutyl lithium to
At the turn of the Millennium Ciba Special- of 3 have been claimed for the treatment imine 5 was developed (Scheme 1). This ap-
ty Chemicals Corporate Research Organisa- of depression and related disorders.[3] The proach is potentially very attractive, but even
tion was pursuing New Business develop- pharmaceutical profiles of 1–3 made these with the best identified catalyst a product
ment projects, related to new production compounds an attractive target for both us with only 40% ee could be obtained,[6] and a
processes for the pharmaceutical industry, and a group at Sepracor.
based on its core competences in the area An efficient route to access any of com- to upgrade the ee to enantiopurity.
of innovative synthesis design, enantio- pounds 1–3 with high enantiopurity allows This drawback led to the development
subsequent resolution was required in order
selective catalysis and control of physical the other compounds to obtain either via of yet another route, where the key step is
properties of actives. During the course of methylation or de-methylation chemistry. the addition of isobutyl lithium to the chi-
these studies we have investigated the enan- Resolution of racemic 1 with a chiral acid ral sulfinimide 6 (Scheme 2). Hydrolysis
tioselective synthesis of sibutramin.
and de-methylation of the resolved 1 with of the resulting product gives (R)-3 with
Racemic sibutramine 1 is one of the diethyl-azodicarboxylate (DEAD) to give excellent yield and enantiopurity.[7] This
few drugs which is licensed for the long- enantiopure 2 was studied at Sepracor.[4] methodology has also recently been applied
term treatment of obesity.[1] On absorption, However, the risks of using DEAD, which for the preparation of active, hydroxylated
the drug is rapidly metabolized to give can violently decompose, ruled this ap- metabolites of sibutramine.[8] However,
the primary metabolites des-methylsibu- proach out for the production of commercial in both cases, the addition is a low-tem-
tramine 2 and di-des-methylsibutramine quantities of 2. As an alternative, a proce- perature reaction (optimum temperature
3 (Fig. 1). Preliminary preclinical studies dure for the resolution of 2 with chiral ac- –78 °C for 6 to 3), and the required auxilia-
suggest that the potent serotonin, norepine- ids was developed, but the efficiency of the ry, which has to be prepared in three steps, is
phrine, and dopamine re-uptake inhibitor resolution was low.[5] Therefore, a catalytic destroyed and thus lost during the workup.
Figꢀ1.
N
NH
NH
2
Cl
Cl
Cl
*Correspondence: Dr. U. Berensa, Dr. A. Hafnerb
aBASF SE
1 Sibutramin
2 des-methylsibutramin
3 di-desmethylsibutramin
GCI/P-M 311
D-67056 Ludwigshafen am Rhein, Germany
Tel.: +49 621 604 8030
E-mail: ulrich.berens@basf.com
bCiba Inc. (now part of BASF)
Performance Chemicals Research
Klybeckstrasse 141, CH-4002 Basel
Tel.: +41 61 636 5437
Schemeꢀ1.
1) DIBAL-H
2) 40% H2NMe
H
i-BuLi, catalyst ee-upgrade
CN
(R)-2
N
40%ee
80%
E-mail: andreas.hafner@ciba.com
Cl
Cl
cSolvias AG
4
5
Klybeckstrasse 191
CH-4002 Basel

60ꢀ CHIMIAꢀ2010,ꢀ64,ꢀNo.ꢀ1/2ꢀ
From ChemiCal researCh to industrial appliCations
Fe, AcOH
1) F3B*OEt2, i-BuLi, -78°C
2) 4N HCl, MeOH
3) NaOH
H2, Rh(I)-cat
Ac2O
H
HN
HN
N
N
+
NH
Cl
O
2
Cl
O
Cl
Cl
S
Cl
OH
-
83%, 99%ee
O
7
9
8
(R)-3
6
Schemeꢀ2.
Schemeꢀ4.
-
O
S
+
TsN
H
Ph
O
Me
Me
ClMg
1.
N
O
iBuMgCl
CN
Cl
+
-
+
S
O
CN
N
TsHN
H
Ph
Cl
THF, -45°C
N
N
2. Ac2O
Cl
MgCl
Cl
Cl
H
Ac
Cl
Ac
Ac
Me
Me
4
10
10a
4
4a
1. HCl
NaOMe
2. NaOH
NaBH /Ti(OiPr)
4
4
3. D-Tartaric acid
-45 to -10°C
HN
H2, cat.
H2, cat.
Cl
+
-
S
O
95:5 dr
TsHN
O
D-TA
NH
N
N
N
2
H
Cl
Cl
H
Ac
Cl
H
Ac
Cl
H
Ac
Me
Me
TsHN
H
OH
Ph
D-TA, 86%, 99% ee
3
Me
11
11a
10
Ph
Me
92%
4b
Schemeꢀ3.
Schemeꢀ5.ꢀSynthesisꢀofꢀdienamideꢀ10
Thus, nitrile 4 was reacted with methal-
lyl magnesium chloride, and the reaction
mixture treated with acetic acid anhydride
(Scheme 5). After de-acetylation of the
intermediate diacetate 10a with sodium
methoxide the novel dienamide 10 was
after recrystallization obtained as the Z-
stereoisomer in 67% yield, whilst the crude
yield of 10 was almost quantitative.
can be hydrogenated very efficiently with
cationic rhodium complexes of DuPHOS-
type ligands.[10a] Such enamides have been
obtained by the reduction of oxime 7 with
iron powder in the presence of acetic acid
and acetic anhydride (Scheme 4).
Very recently, the issue of the recovery
of the chiral auxiliary has been addressed
(Scheme 3).[9] The new route involves
the reaction between the condensation
product of nitrile 4 and i-BuMgCl, and
an optically pure oxathiazolidine-2-oxide
to deliver sulfinate imines 4a. The highly
diastereoselective reduction (95:1 de) of
these imine derivatives to compounds
4b could be optimized to 95:5 de, albeit
at low temperatures again (–45 °C) and
with stoichiometric amounts of the waste-
producing reducing agent NaBH4 in com-
bination with Ti(OiPr)4. Crystallisation of
the d-tartrate salt of the obtained di-des-
methylsibutramine 3 after hydrolysis and
di-methylation finally led to optically pure
sibutramine. On multigram scale, the chiral
auxiliary could be recovered in 92% yield
after hydrolysis of the reduction product.
Defining the stereochemistry of the ste-
reogenic carbon of 3 via enantioselective
hydrogenation of a suitable substrate with
high ee would circumvent the formation of
the undesired enantiomer and thus its loss,
respectively the need to utilize this mate-
rial for example by a racemisation/resolu-
tion process. As mono- or di-methylation
is easy to perform, we devised a new, cata-
lytic asymmetric route to 3, as this provides
also an efficient access to the full range of
our target compounds 1–3.
In our case however, the reduction of
oxime 7 according to the published proto-
col provided 8 only in a maximum 30%
yield. An alternative approach to such ena-
mides involves the reaction of Grignard
reagents with nitriles. However, this ap-
proach usually requires chromatographic
purification of the products and yields are
low.[10b,c] It was thus surprising that the
reaction of isobutyl magnesium chloride
with nitrile 4 and in situ treatment of the
reaction mixture with acetic anhydride
gave enamide 8 in 62% yield. This is a
dramatic yield improvement, as alone for
the synthesis of 7 three additional steps are
required: i) aqueous quench of the Grig-
nard addition reaction, ii) hydrolysis of the
imine with hydrochloric acid to the ketone,
and iii) formation of the oxime 7.
Catalysis
Substrates which are closely related to
10 have been hydrogenated with cationic
rhodium catalysts derived from DuPHOS-
type ligands to provide the g,d-unsaturated
amide with excellent chemo- and enanti-
oselectivity.[12] From the literature, the ex-
pected product from amide 10 should be
11a. Surprisingly, hydrogenation of enam-
ide 10 with rhodium catalysts derived from
the achiral ligands Dipfc and the biaryldi-
phosphine BiPHEP (Fig. 2) gave 8 quanti-
tatively (Table 1). This means that not as
reported the a,b- but the g,d-double bond is
hydrogenated preferentially, which results
in the formation of the inactive substrate
8. Employing the (S,S)-Et-DuPHOS lig-
and under quite harsh conditions furnished
compound 11a in only 25% yield (ee not
determined) together with 9% saturated
amide 9, in 66% ee (R).
Quite surprisingly and in strong con-
trast to reported results,[10a] the cata-
lytic hydrogenation of 8 with cationic
rhodium catalysts such as [Rh-(COD)-
(MeDuPHOS)]BF ,
[Rh-(COD)-(Me-
BPE)]BF4, or [Rh4-(COD)-(Et-Ferrotane)]
BF4 or even with the very active catalyst
[Rh-(COD)-(Dipfc)]BF [11] was extremely
sluggish even under dra4stic conditions. As
the unexpectedly low reactivity of 8 most
probably originates from the severe steric
shielding of the enamide C=C double bond,
a sterically less congested hydrogenation
substrate was sought.
As hydrogenation of enamides with ru-
thenium-based catalysts is well document-
ed,[13] we also tested [RuCl2-(p-cymene)]2
in combination with Biphep, and were de-
lighted to obtain 9 cleanly in quantitative
yield.
Results and Discussion
Compounds such as 8 appear to be suit-
able hydrogenation substrates, as it is well
documented that enamides of this type

From ChemiCal researCh to industrial appliCationsꢀ
CHIMIAꢀ2010,ꢀ64,ꢀNo.ꢀ1/2ꢀ 61
Fig.ꢀ2.ꢀLigandsꢀ
usedꢀinꢀcatalyticꢀ
hydrogenationsꢀ
This prompted us to evaluate the per-
formance of chiral ruthenium catalysts
in the asymmetric hydrogenation of 10,
which were derived from axially chiral
enantiopure ligands such as BINAP or BI-
PHEMP and the like (Table 2).
P(i-Pr)
2
P
P
P
P
P
PPh2
PPh2
Fe
P
P(i-Pr)
2
S
It was a pleasing finding that both sat-
isfactory conversions and high ee could be
accomplished with these ligands. At 50 °C
and 50 bar hydrogen pressure, full conver-
sion to the desired amide 9 in 93.9% ee (R)
was observed using the (R)-BINAP ligand.
These reaction conditions seem to repre-
sent a good compromise for high perform-
ance of the catalyst, since lowering (entry
2) or rising (entry 3) the temperature had
a detrimental effect on catalytic activity
and selectivity, respectively. An analogous
effect of hydrogen pressure on the course
of the reaction was observed (entries 4–5).
Under otherwise similar conditions, other
ruthenium/ BINAP catalysts were found
to be both less active and selective (entries
5–7).
From the testing of several other chiral
diphosphines (entries 9–13), the MeOBi-
PheP ligand provided best performance
with 98.3% ee and full conversion. Inter-
estingly but unexpectedly, enamide 8 is
not fully reduced under the same reaction
conditions, (entry 14).
Hydrogenation of dienamide 10 ap-
pears to produce 11a as the primary prod-
uct, as enamide 8 cannot be hydrogenated
efficiently with the employed ruthenium
catalysts. Quite interestingly, intermediate
11a was rarely detected in concentrations
exceeding ca. 2%, which means that the
remaining double bond is hydrogenated
rather rapidly with a homogeneous catalyst
to give 9. This indicates that presumably an
amide directed homogeneous hydrogena-
tion is in operation. It would be interesting
to study whether this second hydrogena-
tion is also enantioselective by employing
a prochiral analogue of 11a, but no work
towards this end was undertaken.
With these first results at hand, we
further optimized the reaction, employing
BINAP (the cheapest axially chiral diphos-
phine) and MeOBiPheP (most effective for
our transformation at S/C = 100) as ligands
and at higher substrate/catalyst ratios, typi-
cally S/C = 500 (Table 3).
Although the hydrogenation of 10 with
BINAP provides the desired amide 9 in
high ee, at high substrate/catalyst ratios the
conversion was not satisfactory. Whereas
the addition of trifluoroethanol or mixtures
of ethanol with dichloromethane or tet-
rahydrofurane gave no improvement (en-
tries 2–4), it was possible to achieve full
conversion at S/C = 1000, when a small
quantity of hydrochloric acid was added
to the reaction mixture (entry 5). This was
however at the cost of enantioselectivity,
which dropped to 84.7% ee.
(R,R)-Et-Butiphane
Dipfc
BIPHEP
(R,R)-Et-DuPHOS
(R,R)-Me-BPE
P
MeO
MeO
PPh2
PPh2
Me
Me
PPh2
PPh2
PR2
PR2
P
N
H
Fe
R = Ph: (R)-BINAP
R = Tol: (R)-TolBINAP (R)-MeO-Biphep
(R)-Biphemp
Taniaphos
Tableꢀ1.ꢀRh-ꢀandꢀRu-catalysedꢀhydrogenationꢀofꢀenamideꢀ10
Products
pH2
T
9
11
8
Catalyst Precursor
Ligand
S/C Solvent [bar] [°C] [APa %] [APa %] [APa %]
[Rh-(COD)-Dipfc]BF4
–
200
MeOH
MeOH
DCE
5
25
60
50
50
0
<1%
<1%
<1%
25c
100
100
100
66
[Rh-(COD)-Dipfc]BF4
[Rh-(NBD)-Cl]2
–
50
10
20
20
0
BIPHEP
100
100
0
[Rh-(COD)-(S,S)-
EtDuPHOS]BF4
MeOH
9b
[RuCl2-(p-cymene)]2
BIPHEP
100
MeOH
50
50
100
<1%
0
Reactionꢀconditions:ꢀ150–200ꢀmgꢀstartingꢀmaterial;ꢀ[10]ꢀ=ꢀ0.15–0.2ꢀM;ꢀreactionꢀtimeꢀ16–17ꢀh;ꢀ
APꢀ=ꢀHPLCꢀpeaks;ꢀaAreaꢀpercentage;ꢀbeeꢀnotꢀdetermined;cꢀ66% eeꢀ(R).
Tableꢀ2. Catalystꢀscreeningꢀforꢀtheꢀhydrogenationꢀofꢀ10 withꢀchiralꢀrutheniumꢀcatalystsꢀatꢀS/Cꢀ=ꢀ100
Entry Precursor / Ligand
or Precatalyst
pH2
T
Conv.
of
8
9
ee 9
[%]
[bar] [°C]
[AP %] [AP %]
10
[AP %]
1
2
3
[RuCl2(p-cymene)]2ꢀ/ꢀ(R)-BINAP
50
50
50
50
25
80
100
67
0
52
1
100
13
93.9
98.6
88.5
[RuCl2(p-cymene)]2ꢀꢀ/ꢀ(R)-BINAP
[RuCl2(p-cymene)]2ꢀ/ꢀ(R)-BINAP
100
98
4
[RuCl2(p-cymene)]2ꢀ/ꢀ(R)-BINAP
[RuCl2(p-cymene)]2ꢀ/ꢀ(R)-BINAP
RuCl3*3H2Oꢀ/ꢀ(R)-ꢀBINAP
10
90
50
50
50
50
50
50
50
50
50
50
50
50
50
50
100
100
100
100
100
100
100
100
23
25
58
15
10
0
76
75
95.1
92.5
84.9
92.9
91.0
93.7
95.4
98.3
5
6
41
7
[RuCl2-(R)-ꢀBINAP]2ꢀNEt3
84
8
[RuCl-(R)-BINAP-(p-cymene)]Cl
[RuCl2(p-cymene)]2ꢀ/ꢀ(R)-TolBINAP
[RuCl2(p-cymene)]2ꢀ/ꢀ(R)-Biphemp
89
9
100
100
100
10
11
0
[RuCl2(p-cymene)]2ꢀ/ꢀ
(R)-MeOBIPHEP
0
12
13
14a
[RuCl2(p-cymene)]2ꢀ/ꢀ
(S)-(R)-Taniaphos
50
50
50
50
50
50
100
95
–
89
84
40
11
16
60
89.1
20
[RuCl2(p-cymene)]2ꢀ/ꢀ
R,R)-Me-Butiphane
[RuCl2(p-cymene)]2ꢀ/ꢀ
(R)-MeOBIPHEP
85
Reactionꢀconditions:ꢀ150–200ꢀmgꢀstartingꢀmaterial;ꢀ10ꢀ=ꢀ0.15–0.2ꢀMꢀinꢀEtOH;ꢀreactionꢀtimeꢀ16ꢀ
h,ꢀ(R)-9ꢀpreferentiallyꢀformedꢀinꢀeachꢀcase;ꢀaEnamideꢀ8ꢀasꢀstartingꢀmaterial.

62ꢀ CHIMIAꢀ2010,ꢀ64,ꢀNo.ꢀ1/2ꢀ
From ChemiCal researCh to industrial appliCations
Tableꢀ3.ꢀAsymmetricꢀhydrogenationꢀofꢀ10 withꢀchiralꢀrutheniumꢀcatalystsꢀatꢀS/Cꢀ≥ꢀ500
are reported in parts per million (ppm)
relative to external standard TMS (1H,
Conv.
of 10
[AP %]
pH2
[bar] [°C]
T
8
9
ee 9
[%]
13C). Chemicals were ordered from Strem
Entry Ligand
Solvent
[AP %] [AP %]
(for metal complexes and ligands), Fluka
or Aldrich and used as received. Solvents
were ordered from Fluka or Aldrich and
freshly distilled before use under argon
1
ꢀ(R)-BINAP
EtOH
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
80
70
50
50
100
95
65
28
92
94
0
30
72
8
98.0
2
(R)-BINAP
(R)-BINAP
(R)-BINAP
(R)-BINAP
EtOH/DCMꢀ(4:1)
EtOH/THFꢀ(1:4)
TFE
100
100
100
100
100
100
100
62
91.9
atmosphere: EtOH (Mg), CH2Cl and tolu-
ene (CaH2), THF and diethethyl2ether (Na/
3
27
benzophenone). Trifluoroethanol was used
4
6
28
as received from Aldrich. The ruthenium
catalysts were prepared according to.[14]
5a
6b
7
EtOH
100
100
100
42
17
5
84.7
98.5
96.4
92.1
97
Synthesis of 8 via Nitrile Alkylation
Adrythree-necked500mlflaskwithni-
trogen inlet was charged with 1-(4-chloro-
phenyl)-cyclobutanecarbonitrile (4, 20.1 g,
105 mmol) and dry toluene (300 ml). The
mixture was cooled to 5 °C, and then iso-
butyl magnesium bromide (79 ml of a 2M
solution in diethyl ether, 158 mmol) was
added within 15 min. The reaction mixture
was heated to 105 °C, and the diethyl ether
was continuously removed by distillation.
The mixture was kept at reflux (105 °C),
and after 1 h the starting material was con-
sumed completely (TLC). The reaction
mixture was then cooled to 5 °C, and after
theadditionofaceticanhydride(32.1g,315
mmol) the yellow suspension was stirred
at room temperature for another 3 h. The
reaction was quenched with methanol (30
ml), and then neutralized with a saturated
sodium hydrogen carbonate solution (200
ml).After the addition of diethyl ether (300
ml) two layers formed. The organic layer
was washed twice with water, and dried
over sodium sulfate. Evaporation of the
solventinvacuogaveayellow-orangesolid
(35 g), which was recrystallized from hex-
ane to give 8 as pale yellow crystals (19 g,
(R)-MeOBIPHEP EtOH
(R)-MeOBIPHEP EtOH
(R)-MeOBIPHEP i-PrOH
(R)-MeOBIPHEP EtOH
(R)-MeOBIPHEP EtOH
(R)-MeOBIPHEP EtOH
0
0
8
58
25
6
9c
10d
11e
12
n.d.
95.2
100
0
100
Reactionꢀconditions:ꢀ1–2ꢀgꢀstartingꢀmaterial,ꢀmetalꢀprecursor:ꢀ[RuCl2(p-cymene)]2ꢀunlessꢀ
otherwiseꢀnoted,ꢀ[10]ꢀ=ꢀ0.15–0.2ꢀMꢀinꢀsolvent;ꢀreactionꢀtimeꢀ16–26ꢀh,ꢀ(R)-9ꢀpreferentiallyꢀformedꢀ
inꢀeachꢀcase;ꢀa0.4ꢀeq.ꢀHClaqꢀ1Mꢀrelativeꢀtoꢀdienamideꢀ10ꢀwasꢀadded,ꢀS/Cꢀ=ꢀ1000;ꢀb20.0ꢀgꢀscale;ꢀ
c[Ru-(R)-MeOBiphep-(TFA)2]ꢀasꢀcatalyst;ꢀd[Ru-(R)-MeOBiphep-(OAc)2]ꢀasꢀcatalyst;ꢀeS/Cꢀ=ꢀ1500.
In contrast, when the MeOBiPheP
Finally, (R)-di-des-methylsibutramine
ligand was employed, the reaction went 3 could be mono-methylated[15] or dimeth-
smoothly to completion and provided (R)- ylated[16] according to literature proce-
9 quantitatively in excellent 98.5% ee on dures, thus completing the synthesis of
20 g scale (entry 6). Modifying the pres- both (R)-des-methylsibutramine and (R)-
sure, temperature and nature of solvent or sibutramine.
employing isolated ruthenium catalysts[14]
led to erosion of activity and/or selectivity.
At S/C = 1500, it was necessary to rise the Conclusion
temperature to 100 °C in order to achieve
full conversion (entry 11). In this case the
isolated product had 95.2% ee.
1
The first catalytic asymmetric synthesis
62 %). From H-NOE experiments the
of (R)-di-des-methylsibutramine 3, a pre-
product is the (Z)-stereoisomer.
1H NMR (DMSO-D6, 300 MHz): d
0.88 (d, 6 H, 3J = 6.8 Hz, 2 CH3); 1.60–1.90
m (2 H, CH2CH CH2); 1.78 (s, 3 H, CH3);
2.10–2.32 (m, 2 2H, CH2CH CH2); 2.28 (m,
1 H, CH(CH3) ); 2.40–2.502(2 H, CH2CH -
CH2); 5.15 (d2, 1 H, J = 9.6 Hz, C=CH2);
7.21, 7.28 (2 * 2 H, Ar-H); 8.09 (br s, 1
H, NH).
Interestingly, the crystalline solid cursor to the anti-obesity drug sibutramine
amide 9 of high enantiomeric purity 1, has been achieved in 49% overall yield
(>92%) was found to lend itself to an ee- (3 HCl salt) from nitrile 4. The enanti-
.
upgrade by crystallisation. When (R)-9 oselective hydrogenation of intermediate
with an enantiomeric purity of 96.6% was dienamide 10 proved only feasible with
recrystallized from di-isopropyl ether, the ruthenium-based catalysts. Optimization
recovered amide (72%) had an ee of 99.7%. studies have shown that the atropoisomeric
This behavior of 9 is a very useful feature biaryldiphosphine MeOBiPheP is most ef-
of the present synthesis, as the possibility fective for this transformation, bringing
to upgrade the ee allows a freer choice be- excellent conversion and enantioselectivity
tween a variety of hydrogenation catalysts (up to 98.5% ee). The cheap BINAP ligand
13C NMR (DMSO-D6, 75 MHz): d
16.58 (CH2); 23.16 (2 CH3); 23.59 (CH3);
27.39 (CH); 32.76 (2 CH2); 52.13 (C);
128.34, 128.92 (4 Ar CH); 130.83 (Ar
C-Cl); 131.31 (C=CH); 136.83 (C=CH);
146.57 (Ar C); 168.46 (C=O).
respectively the corresponding ligands.
proved also interesting in particular in view
The de-acetylation of 9 to give 3 was of large-scale applications and for overall
successfully accomplished with HCl at process cost reasons (although further op-
elevated temperature. Surprisingly, even timization would be needed). We anticipate
under the harsh conditions required for the that this synthetic route should be amenable
cleavage, the enantiomeric excess of the to the large-scale production of optically
starting material is retained in the product. pure sibutramine 1 and derivatives thereof.
When a material with 95.1% ee (R) was hy-
Synthesis of (Z)-N-{1-[1-(4-Chloro-
phenyl)-cyclobutyl]-3-methyl-buta-
1,3-dienyl}-acetamide 10
A dry 500 ml three-necked flask with
nitrogen inlet was charged with 1-(4-chlo-
drolyzed and re-acetylated under standard
conditions (Ac O, DMAP cat., CH2Cl2, r. t., Experimental
ro-phenyl)-cyclobutanecarbonitrile
(4,
3 hrs, see Expe2rimental section), (R)-9 was
20.1 g, 105 mmol) and dry THF (300 ml).
The mixture was cooled to 5 °C, and me-
thallyl magnesium chloride (105 ml of a
obtained back in 67% ee and 95.7% ee (no
direct assay for 3 was available).
NMR spectra were recorded on a Vari-
an 400 MHz spectrometer. Chemical shifts

From ChemiCal researCh to industrial appliCationsꢀ
CHIMIAꢀ2010,ꢀ64,ꢀNo.ꢀ1/2ꢀ 63
freshly prepared solution 1.5 M in THF, ring was started. After 17 h the autoclave 32.18 (CH2CH2CH2); 40.22 (CH2); 50.46
158 mmol, 1.5 eq.) was added within 30 was cooled to ambient temperature and the (q C);53.17 (CNHAc); 128.13, 128.64
min. The reaction mixture was stirred for pressure released. The resulting pale yel- (Aryl CH); 132.01 (q, CCl); 144.39 (q,
another 30 min at 5 °C, and then slowly low solution was evaporated under reduced 4-ClPhC); 169.91 (COCH3).
warmed to room temperature, before ace- pressure (rotavapor, max bath T/°C = 40) to
tic anhydride (315 ml of a 1 M solution give the product mixture which was ana- Hydrogenation of 10 with the Ru-
in THF, 315 mmol, 3 eq.) was added. The lyzed using the following assay.
orange reaction mixture was stirred at 60
thenium/(R)-MeOBiPheP Catalyst
on Large-scale at S/C = 500.
A 50 ml Schlenk flask equipped with
a magnetic stirring bar was charged with
[Ru-CI -(p-cymene)] (42.3 mg, 69.0 mmol)
°C until the acetylation was complete (2–4 Analytical Assay for the Hydro-
h, monitored by TLC). This gave a mix- genation Products and Starting
ture, which contained both 10 and the N- Materials
di-acetylated product (ca. 1:2 ratio). The
The following HPLC-method was used and (R)2-MeOBiPheP2(80.4 mg, 138 mmol),
excess of acetic anhydride was quenched for the determination of the ee of the prod- evacuated under high vacuum/argon flush-
with 20 ml methanol, and after the addition ucts. For the assignment of the absolute ing (this operation was repeated three times)
of sodium methylate (160 g of a 15 % solu- stereochemistry of 3 a small quantity of and then charged under stirring with dry, de-
tion in methanol, 445 mmol), and further this material was prepared as described in gassed EtOH (20 ml). Dienamide10(20.0g,
stirring for 15 min the by-product 11a had the literature.[7]
been de-acetylated to give 10. The reaction
69.0 mmol) was taken into a 300 ml auto-
Column: Chiracel OD-H, diam. = 0.46 clave and set under argon. The catalyst solu-
mixture was then diluted with ethylacetate cm, length: 25 cm, mobile phase hexane/ tion was transferred via canula under argon
(250 ml), and then washed with saturated EtOH 98:2, flow: 0.7 ml/min, temperature: atmosphere to the autoclave. Then 130 ml
ammonium chloride (500 ml), brine (500 25 °C, detection 230 nm, injection volume: of dry, degassed EtOH was transferred
ml), and water (500 ml). The organic layer 1.0 ml, sample preparation: ca. 5 mg in 1 to the autoclave and the resulting mixture
was dried (sodium sulfate), and removal of ml hexane/ethanol (4:1). Retention times: submitted to hydrogen pressure (10 bar) and
the solvent in vacuo gave 10 (30 g) as beige di-enamide 10: 23.6 min, mono-enamide the pressure released. After three cycles, the
crystals. Recrystallization from di-isopro- 8: 17.2 min, enantiomers of amide 11: pressure was set to 50 bar and the tempera-
pyl ether furnished pure 10 (20 g, 67 %) 14.5 and 15.5 min, amide (S)-9: 12.9 min, ture to 50 °C; 30 min later, magnetic stir-
1
as pale beige crystals. From H-NOE ex- amide (R)-9: 13.9 min.
ring was started. After 26 h, an 1H-NMR of
a reaction aliquot showed complete conver-
sion to the desired product. The pressure was
released, the autoclave set under argon and
periments the major stereoisomer is the
(Z)-stereoisomer, mp = 120 °C.
Enantioselective Hydrogenation of
1H NMR (DMSO-D6, 300 MHz): d 10 at S/C = 1500
1.63–1.74 (m, 2 H, CH ); 1.74 (s (br), 3
A 10 ml Schlenk flask equipped with the pale yellow solution evaporated under
H, CH3); 1.78, (s, 3 H,2 CH3C=O); 2.28, a magnetic stirring bar was charged with reduced pressure (rotavapor, max bath T/°C
2.49 (2 m, 2 H each, 2 CH ); 4.81 (s, 1H, [Ru-Cl2-(p-cymene)] (1.40 mg, 2.3 mmol) = 40) to give amide 9 in quantitative yield
Z-C=CH2); 4.91 (s, 1 H, E2-C=CH2); 5.95 and (R)-MeOBiPheP2(2.80 mg, 4.8 mmol), (19.9 g) and 98.5% ee (R). [a]25 = +3.16
D
(s, 1 H, C=CH); 7.26, 7.30 (m, 2 H each, evacuated under high vacuum/argon flush- (c = 10.065 in CHCl3).
H-aryl); 8.34 (s (br), NH).
ing (this operation was repeated three
13C NMR (DMSO-D6, 75 MHz): times) and EtOH (3 ml) added with stir- Hydrolysis of Amide 9 to Di-des-
16.57 (CH2); 21.39 (CH ); 23.53 ring. Dienamide 10 (2.00 g, 6.90 mmol) methylsibutramin 3
d
(CH C=O); 33.11 (2 CH2); 533.33 (C), was taken into a 25 ml Schlenk flask, set
A 86 ml tantalum autoclave equipped
117.393 (C=CH2); 125.77 (C=CH); 128.43, under argon and dissolved in ethanol (13 with a Teflon stirring bar was charged with
129.11(4Ar CH); 130.99 (Ar C-Cl) 139.80 ml). The catalyst and starting material solu- 9 (1.0 g, 3.4mmol, 95.1% ee) and hydrochlo-
(NC=C); 141.42 (C=CH2); 146.35 (Ar C); tions were transferred sequentially to a 50 ric acid (50 ml, 37% in water, 185 mmol). The
169.22 (C=O).
ml thermostated stainless steel autoclave autoclave was closed and heating at 180 °C
equipped with a magnetic stirring bar, was started. After 90 min the internal tem-
under argon atmosphere. The autoclave perature had reached 180 °C at a pressure of
was submitted to hydrogen pressure (10 44 bar. After 9 h, the heating was stopped,
bar) and the pressure released. After three and the reaction mixture cooled down to
Hydrogenation Screening of 10
with Achiral Rhodium and Ruthe-
nium Catalysts
Typical procedure: To a 10 ml Schlenk cycles, the pressure was set to 50 bar and roomtemperaturewithin11h(atthatpointthe
flask with a magnetic stirring bar was the temperature to 100 °C; 20 min later, internal pressure was 3 bar). The pressure was
charged the respective catalyst. The Sch- magnetic stirring was started. After 17 h, released, the autoclave opened and the beige
lenk flask was evacuated and flushed with the pressure was released and the resulting reaction mixture taken out (a small amount
argon three times. Then the degassed sol- pale yellow solution evaporated under re- of a black thick oil had also formed on the
vent (3 ml) was added, and the catalyst duced pressure (rotavapor, max bath T/°C upper wall). The autoclave was rinsed with
dissolved. The substrate 10 was trans- = 40) to give amide 9 in quantitative yield distilled water (3 × 10 ml) and evaporation
ferred into a 25 ml Schlenk flask, which and 95.2% ee (R).
of the reaction mixture to dryness (rotavapor
was purged by three cycles vacuum/argon
1H-NMR (CDCl , 400 MHz): d 0.59– and high vacuum at 60 °C, 1h) gave 725 mg of
flushing, and then dissolved in the solvent 0.71, 1.16–1.26 (23 m, 1 H each, CH 3 (HCI salt) as a beige solid (73%).
(3 ml). The solution of both the catalyst CHMe2); 0.82 (d, 3 H, J = 6.6 Hz, CH3)2;
1H NMR (DMSO-D6, 300 MHz): d 0.8
and the substrate were transferred sequen- 0.95 (d, 3 H, J = 6.6 Hz, CH3); 1.39–1.52 (d, 3 H, CH3); 0.85 (d, 3 H, CH ); 0.9, 1.15
tially into a 50 ml thermostated stainless (m, 1 H, CHMe ); 1.75–1.88 (m, 1 H), (2M, 2H each, CH CH CH2); 13.6–1.65 (m,
steel autoclave, which was equipped with 2.05–2.42 (m, 5 2H) (CH CH2CH2); 1.99 2 H, CHCH2); 1.92(m,21 H, CHCH ); 2.3,
a magnetic stirring bar under an argon at- (s, 3 H, COCH ); 4.52 (m2, 1 H, CHNAc); 2.5 (2M, 2H each, CH2CH2CH2); 3.245 (m,
mosphere. The autoclave was submitted to 4.76 (br d, 1 H,3NH); 7.04, 7.29 (2 m, 2 H 1 H, CHN); 7.35, 7.45 (m, 2 H each, H-
hydrogen pressure (10 bar) and the pres- each, aryl H).
sure released. After three cycles, the pres-
aryl); 7.95 (s (br), H3N+).
13C-NMR (CDCl3, 100 MHz) d 15.93
13C NMR (DMSO-D6, 75 MHz) d 15.2
sure and temperature were set to the de- (CH CH2CH ); 22.12 (CH3); 23.96 (CH (CH ); 21.4 (CH ); 23.5 (CH3); 24.0 (CH);
sired level, and 20 min later magnetic stir- CO);2 24.29 (2CH3); 25.29 (CHMe2); 31.893, 32.027 and 32.143(2 CH2); 37.7 (CH2); 48.7

64ꢀ CHIMIAꢀ2010,ꢀ64,ꢀNo.ꢀ1/2ꢀ
From ChemiCal researCh to industrial appliCations
Table 4. Enhancement of the enantiomeric purity of acetamide 9 by recrystallisation
di-isoproyl-
ether [ml]
recryst. 9
[g] / [%]/
ee recryst.
ee motherl.
Amide 9 [g]
ee [%]
[%]
[%]
1.5
2.5
92.7
96.6
80
1.3 / 86.6
1.8 / 72.0
92.2
99.7
nd
130
89.0
[3] a) J. W.Young, PCT Application WO 94/00047;
b) J. W.Young, PCT Application WO 94/00114.
[4] a) Q. K. Fang, C. H. Senanayake, Z. Han, C.
Morency, P. Grover, R. E. Malone, H. Bulter, S.
A. Wald, S. T. Cameron, Tetrahedron: Asymm.
1999, 10, 4477; b) Z. Han, D. Krishnamurthy,
K. Q. Fang, S. A. Wald, C. H. Senanayake,
Tetrahedron: Asymmetry 2003, 14, 3553.
[5] Z. Han, D. Krishnamurthy, D. Pflum, Q. K.
Fang, H. Butler, S. T. Cameron, S. A. Wald, C.
H. Senanayake, Tetrahedron: Asymm. 2002, 13,
107.
[6] D. Krishnamurthy, Z. Han, S. A. Wald, C. H.
Senanayake, Tetrahedron Lett. 2002, 43, 2331.
[7] Z. Han, D. Krishnamurthy, D. Pflum, P. Grover,
S. A. Wald, C. H. Senanayake, Org. Lett. 2002,
4, 4025.
[8] B. Z. Lu, C. Senananyake, N. Li, Z. Han, R. P.
Bakale, S. A. Wald, Org. Lett. 2005, 7, 2599.
[9] Z. Han, D. Krishnamurthy, C. H. Senananyake,
Org. Proc. Res. Dev. 2006, 10, 327.
[10] a) M. J. Burk, G. Casy, N. B. Johnson, J. Org.
Chem. 1998, 63, 6084; b) Y. Heng-Suen, A.
Horeau, H. B. Kagan, Bull. Soc. Chem. Fr.
1965, 1454; c) H. B. Kagan, N. Langlois, T. P.
Dang, J. Organomet. Chem. 1975, 90, 353.
[11] M. J. Burk, T. G. P. Harper, J. R. Lee, C.
Kalberg, Tetrahedron Lett. 1994, 35, 4963.
[12] M. J. Burk, J. G. Allen, W. F. Kiesman, J. Am.
Chem. Soc. 1998, 120, 657.
[13] a) M. Kitamura, M. Yoshimura, M. Tsukamoto,
R. Noyori, Enantiomer 1996, 1, 281; b) H.-
U. Blaser, C. Malan, B. Pugin, F. Spindler, H.
Steiner, M. Studer, Adv. Synth. Catal. 2003,
345, 103.
[14] B. Heiser, E. A. Broger, Y. Crameri, Tetrahed-
ron: Asymm. 1991, 2, 51.
[15] S. Krishnamurthy, Tetrahedron Lett. 1982, 23,
3315.
[16] J. E. Jeffery, F. Kerrigan, T. K. Miller, G. J.
Smith, G. B. Tometzki, J. Chem. Soc. Perkin
Trans. 1, 1996, 21, 2583.
(C), 55.9 (C-N), 128.3 and 130.0 (4 Ar
CH); 131.6 (Ar C-Cl); 142.3 (Ar C).
Analysis of the ee of 3·HCl:
An aliquot from the above de-acylation
reaction mixture (630 mg of crude 3·HCl)
wasaddedtoanaqueoussolutionofsodium
hydroxide (10 ml, 1M in water, 10 mmol)
and the mixture was extracted with dichlo-
romethane (3 × 10 ml). The organic phase
was dried (sodium sulfate), and directly
taken into a 100 ml flask with a stirring
bar. Acetic acid anhydride (0.40 ml, 4.2
mmol) was added dropwise with stirring
at room temperature, followed by DMAP
(51 mg, 0.42 mmol). After 3 h, the reaction
mixture was evaporated to dryness. The ee
of this material was 95.7%, which is within
experimental error identical to the ee of the
starting material 9 (95.1% ee).
The crude brown solid was then puri-
fied by chromatography on silica gel (elu-
ent: dichloromethane/methanol 95:5) to
give 580 mg of acetamide 9 as a beige sol-
id. The yield for the sequence hydrolysis/
re-acetylation was 67%, corrected for the
aliquot which was used for the re-acetyla-
tion (Table 4).
Received: December 18, 2009
[1] A. Wirth, J. Krause, J. Am. Med. Assoc. 2001,
286, 1331
[2] T. Jerussi, C. H. Senanayake, Q. K. Fang, PCT
Application WO 00/10551.