A ring-closing metathesis-based approach to the synthesis of (+)-tetrabenazine

Article abstract of DOI:10.1021/ol301612q

A modular stereoselective synthesis of the vesicular monoamine transport inhibitors (+)-tetrabenazine ((+)-1) and (+)-α-dihydrotetrabenazine ((+)-2) has been developed. The approach is based on amine 4 and acid 5 as the key building blocks, which were elaborated into macrolactam 3 by amide coupling and a subsequent highly E-selective RCM reaction. Macrolactam 3 could be converted into tetrabenazine in three known steps.

Full text of DOI:10.1021/ol301612q

ORGANIC
LETTERS
2012
Vol. 14, No. 14
3752–3755
A Ring-Closing Metathesis-Based
Approach to the Synthesis of
(þ)-Tetrabenazine
Manuel Johannes and Karl-Heinz Altmann*
Swiss Federal Institute of Technology (ETH) Zu€rich, HCI H405, Wolfgang-Pauli-Strasse 10,
CH-8093 Zu€rich, Switzerland
karl-heinz.altmann@pharma.ethz.ch
Received June 12, 2012
ABSTRACT
A modular stereoselective synthesis of the vesicular monoamine transport inhibitors (þ)-tetrabenazine ((þ)-1) and (þ)-R-dihydrotetrabenazine
((þ)-2) has been developed. The approach is based on amine 4 and acid 5 as the key building blocks, which were elaborated into macrolactam 3 by
amide coupling and a subsequent highly E-selective RCM reaction. Macrolactam 3 could be converted into tetrabenazine in three known steps.
Tetrabenazine (TBZ, 1; Figure 1) is a synthetic benzo-
quinolizine derivative and as such is structurally related to
alkaloids of the protoberberine and the ipecac families.1
TBZ was in clinical use for the treatment of neuroses and
psychoses in several European countries in the early
1960s,2 before it was withdrawn from the market in
1966.3 Since 2008, racemic TBZ is the only drug approved
by the US FDA for the treatment of chorea, one of
the most common motor dysfunctions associated with
Huntington’s disease (HD);3,4 it is also approved for the
treatment of different movement disorders in a number of
other countries.
TBZ reversibly binds to and inhibits the function of the
vesicular monoamine transporter 2 (VMAT2),5 which is
responsible for monoamine transport from the cytoplasm
into granular vesicles of presynaptic neurons. Inhibition of
VMAT2 leads to enhanced cytoplasmic degradation of
monoamine neurotransmitters by monoamine oxidases,
and the resulting depletion of monoamines from nerve
terminals is believed to underlie the beneficial effects of
TBZ on hyperkinetic movement disorders, including
Huntington’s chorea.3
In vivo, TBZ undergoes rapid and extensive hepatic
metabolism to the corresponding 2-hydroxy derivatives
(dihydrotetrabenazines, Figure 1) as the pharmacologi-
cally active species.3,4 Importantly, the binding of dihy-
drotetrabenazines to VMAT2 is highly stereospecific, with
Ki values of 3.96 nM and 13.4 nM for the R- and β-dihydro
derivatives 2R and 2β, respectively, which are derived from
(þ)-tetrabenazine ((þ)-1);6,7 in contrast, binding affinities
for the corresponding reduction products derived from
(ꢀ)-1 are only in the micromolar range.6,7 For TBZ itself,
the literature data on stereospecific binding (or lack
thereof) are contradictory.6,8
(1) Brossi, A.; Lindlar, H.; Walter, M.; Schnider, O. Helv. Chim. Acta
1958, 41, 119. See also: Pletscher, A. Science 1957, 126, 507.
(2) Pletscher, A.; Brossi, A.; Gey, K. F. Int. Rev. Neurobiol. 1962, 4,
275.
(3) For a recent review on tetrabenazine, see: Guay, D. R. P. Am. J.
Geriatr. Pharmacother. 2010, 8, 331.
(4) US Food and Drug Administration, Center for Drug Evaluation
and Research. New Drug Application No. 21-894: Tetrabenazine.
Medical and Clinical Pharmacology reviews.
(5) (a) Scherman, D.; Henry, J. P. Biochem. Pharmacol. 1980, 29,
1883.
Figure 1. Structures of (þ)-TBZ ((þ)-1) and dihydro-TBZ (2).
r
10.1021/ol301612q
Published on Web 06/28/2012
2012 American Chemical Society

In light of the significant difference in VMAT2 inhibi-
tion between dihydrotetrabenazines derived from (þ)- or
(ꢀ)-1, stereoselective access to (þ)-1 is of considerable
interest, alsointhe context ofdeveloping VMAT2-directed
imaging agents.9 So far, two asymmetric syntheses of (þ)-1
have been reported in the literature,10,11 both of which
depart from 6,7-dimethoxy-3,4-dihydroisoquinoline and
deliver (þ)-1 in eight steps for the longest linear sequence
(in both cases).12
(þ)-1 that would involve the separate installation of the
stereocenter at C3 and that we felt could be adapted to the
incorporation of substituents at the 6 and/or 7 position(s)
or to changes in the size of the central ring more readily
than the existing approaches.14
As illustrated in Scheme 1, the synthesis was to proceed
through macrolactam 3, whose acid-catalyzed cyclization
to the quinolizine scaffold of (þ)-1 had been previously
described by Suh and co-workers.10 In contrast to this
previous work, however, macrolactam 3 would not be the
product of an intramolecular rearrangement but was en-
visaged to be formed through ring-closing olefin metath-
esis (RCM) between C1 and C11b (tetrabenazine
numbering). The requisite diene would be obtained by
amide bond formation between amine 4 and acid 5; amine
4 would derive from an appropriately protected iodide 6 by
way of Stille coupling with Bu3SnCHdCH2, while 5 was to
be obtained via the stereoselective aldol reaction of the
Evans oxazolidinone 7 and acrolein (8).
Scheme 1. Retrosynthesis of (þ)-Tetrabenazine ((þ)-1)
On the basis of previous literature reports,15 initial
attempts at carboxylic acid 5 involved reaction of the boron
enolate of 7 with acrolein (8). However, while the desired
aldol product was indeed formed under these conditions,
the reaction proved to be poorly reproducible, with isolated
yields ranging from 30 to 65%. As an alternative, the use of
the Ti-enolate of 7 was investigated according to methodol-
ogy that has been developed by Crimmins and co-
workers.16 Thus, treatment of 7 with TiCl4 as a Lewis acid
and 1 equiv of N-methyl-2-pyrrolidone as an additive at
ꢀ78 °C followed by the addition of acrolein resulted in the
formation of the desired 2S,3R-aldol product 9 in a repro-
ducible fashion and in good yield (78%; Scheme 2). One
other isomer was detectable by TLC, but this compound
could be cleanly removed by flash chromatography (FC)
and was not further characterized.
One of the objectives of our own work on tetrabenazine
is the assessment of the VMAT-inhibitory activity of
analogues with modifications in the central tetrahydropyr-
idine ring, as no SAR information for this part of the
tetrabenazine structure is available so far.13 In this context,
we have investigated an alternative, modular approach to
(6) Yao, Z.; Wei, X.; Wu, X.; Katz, J. L.; Kopajtic, T.; Greig, N. H.;
Sun, H. Eur. J. Med. Chem. 2011, 46, 1841.
(7) See also: (a) Kilbourn, M. R.; Lee, L. C.; Heeg, M. J.; Jewett,
D. M. Chirality 1997, 9, 59. (b) Kilbourn, M.; Lee, L.; Vander Borght,
T.; Jewett, D.; Frey, K. Eur. J. Pharmacol. 1995, 278, 249.
(8) Yu, Q-s.; Luo, W.; Deschamps, J.; Holloway, H. W.; Kopajtic, T.;
Katz, J. L.; Brossi, A.; Greig, N. H. ACS Med. Chem. Lett 2010, 1, 105.
(9) See, e.g.: (a) Malaisse, W. J.; Louchami, K.; Sener, A. Nature
Rev., Endocrin. 2009, 5, 394. (b) Ravina, B.; Eidelberg, D.; Ahlskog,
J. E.; Albin, R. L.; Brooks, D. J.; Carbon, M.; Dhawan, V.; Feigin, A.;
Fahn, S.; Guttman, M.; Gwinn-Hardy, K.; McFarland, H.; Innis, R.;
Katz, R. G.; Kieburtz, K.; Kish, S. J.; Lange, N.; Langston, J. W.;
Marek, K.; Morin, L.; Moy, C.; Murphy, D.; Oertel, W. H.; Oliver, G.;
Palesch, Y.; Powers, W.; Seibyl, J.; Sethi, K. D.; Shults, C. W.; Sheehy,
P.; Stoessl, A. J.; Holloway, R. Neurology 2005, 64, 208.
Scheme 2. Synthesis of Carboxylic Acid 5
(10) Paek, S.-M.; Kim, N.-J.; Shin, D.; Jung, J.-K.; Jung, J.-W.;
Chang, D.-J.; Moon, H.; Suh, Y.-G. Chem.;Eur. J. 2010, 16, 4623.
(11) Rishel, M. J.; Amarasinghe, K. K. D.; Dinn, S. R.; Johnson,
B. F. J. Org. Chem. 2009, 74, 4001.
(12) For a recent 10-step racemic synthesis of 1, see: Son, Y. W.;
Kwon, T. H.; Lee, J. K.; Pae, A. N.; Lee, J. Y.; Cho, Y. S.; Min, S.-J. Org.
Lett. 2011, 13, 6500.
(13) For reviews on SAR studies on TBZ see: (a) Wimalasena, K.
Med. Res. Rev. 2011, 31, 483. (b) Zheng, G.; Dwoskin, L. P.; Crooks,
P. A. AAPS J. 2006, 8, E682.
The anticipated and desired R-configuration at C3
of aldol product 9 was unequivocally established by
Mosher ester analysis.17,18 Reductive cleavage of the chiral
(14) For both exisiting syntheses of (þ)-1 the stereocenter at C3 is
installed in an intramolecular fashion (either by cyclization or
rearrangement), with the corresponding precursor already incorporat-
ing a stereocenter at the position corresponding to C11b in 1. It is well
conceivable that the stereochemical outcome of these intramolecular
steps would be affected by remote stereocenters at postions 6 and/or 7,
but this would have to be established experimentally.
(15) Gierasch, T. M.; Chytil, M.; Didiuk, M. T.; Park, J. Y.; Urban,
J. J.; Nolan, S. P.; Verdine, G. L. Org. Lett. 2000, 2, 3999.
(16) Crimmins, M. T.; She, J. Synlett 2004, 8, 1371.
(17) (a) Hoye, T. R.; Jeffrey, C. S.; Shao, F. Nature Protocols 2007, 2,
2451. (b) Dale, J. A.; Mosher, H. S. J. Am. Chem. Soc. 1973, 95, 512.
(c) Dale, J. A.; Dull, D. L.; Mosher, H. S. J. Org. Chem. 1969, 34, 2543.
Org. Lett., Vol. 14, No. 14, 2012
3753

auxiliary in 9 with LiBH4 gave a 1,3-diol that was converted
into the corresponding acetonide by treatment with 2,2-
dimethoxypropane and camphorsulfonic acid; J coupling
constant analysis for this cyclic derivative confirmed the syn
configuration of the C2 and C3 substituents in aldol
product 9.18 TBS protection of the secondary hydroxyl
group in 9 then turned out to be more challenging than
expected. Thus, the reaction of 9 with TBSCl and imidazole
proceeded very slowly and gave the desired TBS ether only
in low yield (37%). Rather surprisingly, the yield obtained
upon treatment of 9 with TBSOTf and 2,6-lutidine was
even lower (9%); in addition, yields were not reproducible
for either of these approaches. As a consequence, imide 9
was first converted into β-hydroxy acid 10 in a two-step
sequence that involved cleavage of the auxiliary with benzyl
mercaptanandn-BuLi (to form the corresponding thioester
in 76% yield) followed by saponification of the thioester
with LiOOH in MeOH/THF, which produced 10 in ex-
cellent yield (89%).15 Reaction of 10 with 6 equiv of
TBSOTf in the presence of 12 equiv of 2,6-lutidine gave a
bis-silylated intermediate,19 which upon treatment with
K2CO3 in THF/MeOH/water (after aqueous workup of
the silylation reaction) gave the desired building block 5 in
89% yield after FC (Scheme 2). It should be noted here that
the attempted TBS protection of the intermittent thioester
with TBSCl/imidazole was again unsuccessful, contrary to
what has been reported in the literature.15
The N-BOC derivative 6a was prepared from known
iodide 1120 by carbamoylation of the amide nitrogen with
BOC2O followed by selective cleavage of the acetamide
moiety with hydrazine.21 Stille coupling between 6a and
Bu3SnCHdCH2 in the presence of Pd2(dba)3 and Ph3As
then proceeded smoothly and provided the desired cou-
pling product 12 in 88% yield. The efficiency of the
coupling reaction was strongly dependent on the nature
of the palladium ligand employed, with Ph3As22 leading to
significantly enhanced yields over Ph3P, which was used in
initial experiments. Unfortunately, the only product obtained
from 12 under standard acidic BOC cleavage conditions was
1,2,3,4-tetrahydro-6,7-dimethoxy-1-methylisoquinoline,
while none of the desired amine 4 could be isolated. The
transformation of 12 into 4 could eventually be accom-
plished by base-mediated removal of the BOC group,23
albeit in very moderate yield of 52%.
In light of the low overall yield for the elaboration of
iodide 11 into amine 4 we investigated the use of FMOC-
protected amine 13 (obtained by reaction of β-(3,4-
dimethoxyphenyl)ethylamine with Fmoc-OSu in 91%
yield) as a possible alternative starting material for the
preparation of 4. Treatment of 13 with iodine chloride in
glacial acetic acid gave iodide 6b in excellent yield (96%);
subsequent Stille coupling with Bu3SnCHdCH2 resulted
inconcomitantlossoftheFmocgroup and providedamine
4 in 53% overall yield for the 4-step sequence from β-(3,4-
dimethoxyphenyl)ethylamine, thus establishing efficient
access to this building block.
The subsequent coupling of 4 with carboxylic acid 5 was
accomplished with 1-ethyl-3-(3-dimethylaminopropyl)-
carbodiimide (EDCI) in the presence of 1-hydroxy-7-
azabenzotriazole (AtOH) in DMF24 and gave the desired
diene 14 in good yield (72%).
Scheme 3. Synthesis of Amine 4 and Diene 14
While the strategy described above eventually provided
satisfactory access to diene 14, it had not been clear,
initially, if the use of intermediate 6b would in fact lead
to an improved synthesis of amine 4. Thus, we also
investigated an alternative approach to 14, where the
critical Stille coupling would be performed only after
amide bond formation. As illustrated in Scheme 4, this
strategy was based on the EDCI/AtOH-mediated coupling
of acid 5 with amine 16, which provided amide 17 in high
yield. The latter underwent Stille coupling with
Bu3SnCHdCH2 (Pd2(dba)3, Ph3As) in a highly efficient
manner to furnish diene 14 in 88% yield.
The requisite amine 16 was prepared in two steps from
nitrile 15, which underwent regioselective iodination with
iodine chloride in acetic acid (Scheme 4);19 the resulting
crystalline o-iodobenzylcyanide could then be reduced
For the synthesis of amine 4, the two differently N-pro-
tected iodides 6a and 6b were explored as substrates for the
crucial Stille coupling with Bu3SnCHdCH2 that serves to
install one of the two terminal alkene moieties of the RCM
precursor (Scheme 3).
(20) Lete, E.; Collado, M. I.; Sotomayor, N.; Vicente, T.; Villa, M.-J.
J. Heterocycl. Chem. 1995, 32, 1751.
(21) The efficient direct iodination of N-BOC-β-(3,4-dimethoxyphenyl)
ethylamine with CF3C(O)OAg/I2 has been described recently: Sawant,
R. T.; Waghmode, S. B. Synth. Commun. 2011, 41, 2385.
(22) Farina, V.; Krishnan, B. J. Am. Chem. Soc. 1991, 113, 9585.
(23) Tom, N. J.; Simon, W. M.; Frost, H. N.; Ewing, M. Tetrahedron
Lett. 2004, 45, 905.
(18) For NMR data, see the Supporting Information.
(19) Boger, D. L.; Patane, M. A.; Zhou, J. J. Am. Chem. Soc. 1994,
116, 8544.
(24) Yang, D.; Zhang, Y.-H.; Li, B.; Zhang, D.-W. J. Org. Chem.
2004, 69, 7577.
3754
Org. Lett., Vol. 14, No. 14, 2012

Scheme 4. Alternative Synthesis of Diene 14
Scheme 5. Ring-Closing Metathesis of 14
with borane in THF,25 providing amine 16 in moderate,
but still acceptable, yield (55%). Compared to the synthe-
sis of diene 14 by the direct coupling of acid 5 and amine 4,
this alternative approach requires one additional step in
the longest linear sequence (from 7; seven vs six steps). This
is reflected in a lower overall yield of 27% (from 7) vs 30%
for the route via amine 4.
With diene 14 in hand, the stage was set for the explora-
tion of the key metathesis reaction.26 First attempts to
accomplish RCM-based macrocyclization of 14 were car-
ried out with the Grubbs II catalyst27 in refluxing toluene.
However, under these conditions none of the desired
macrolactam was obtained; instead, dimer 18 (Scheme 5)
was isolated in 41% yield together with traces of the
stilbene that arises from cross metathesis of 14 with the
benzylidene moiety of the catalyst.
No reaction was observed in dichloromethane or di-
chloroethane, with starting material being recovered un-
changed, independent of the reaction temperature (rt or
reflux). Gratifyingly, the use of the more reactive Hoveydaꢀ
Grubbs II catalyst28 in refluxing toluene afforded the
desired E configured macrolactam 3 in excellent yield,
although the compound could only be isolated as a 10:1
mixture with dimer 18 (83%). The latter could not be
chromatographically removed at this stage (Scheme 5).
At the stage of macrolactam 3 our strategy toward (þ)-1
converged with the synthesis that had been previously
developed by Suh and co-workers.10 Thus, acid-induced
ring closure of 3 with concomitant cleavage of the TBS ether
gave 4-oxo benzoquinolizine 19 in 68% yield (based on the
10/1 mixture of 3and 18, Scheme 6); the deprotected form of
dimer 18 from the RCM step could be readily removed at
this stage. LAH reduction of 19 followed by TPAP oxida-
tion then provided (þ)-1 in 29% overall yield from 3
(corrected for the purity of 3; vs 42% reported by Suh).
Scheme 6. Final Steps in the Synthesis of (þ)-Tetrabenazine10
In summary, we have established a new modular asym-
metric route to (þ)-tetrabenazine ((þ)-1) that is based on the
synthesis of key intermediate 3 by ring-closing metathesis.
Lactam 3 was obtained in seven steps and 27% overall yield
from acyl oxazolidinone 7 (i.e., with an average per step yield
of 81%). Our approach to 3is somewhat less efficient than the
one elaborated by Suh and co-workers (four steps and 54%
overall yield from 6,7-dimethoxy-3,4-dihydroisoquinoline);
however, because of its modularity it should be more readily
adaptable to the synthesis of tetrabenazine analogues with
modifications of the central ring. Studies along these lines are
currently ongoing in our laboratory.
Acknowledgment. This work was supported by the
Swiss National Science Foundation (NCCR TransCure).
We thank Dr. Bernhard Pfeiffer (ETHZ) and Fabienne
Gaugaz (ETHZ) for NMR support and Louis Bertschi
from the ETHZ-LOC MS-Service for HRMS spectra.
Supporting Information Available. Synthetic proce-
1
dures, complete spectroscopic data, H- and 13C NMR
spectra for all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org
(25) Fowler, J. S.; MacGregor, R. R.; Wolf, A. P. J. Med. Chem.
1976, 19, 356.
(26) For recent reviews on the metathesis reaction, see: (a) Hoveyda,
A. H.; Zhugralin, A. R. Nature 2007, 450, 243. (b) Nicolaou, K. C.;
Bulger, P. G.; Sarlah, D. Angew. Chem., Int. Ed. 2005, 44, 4490.
(27) Scholl, M.; Trnka, T. M.; Morgan, J. P.; Grubbs, R. H. Tetra-
hedron Lett. 1999, 40, 2247.
(28) Garber, S. B.; Kingsbury, J. S.; Gray, B. L.; Hoveyda, A. H.
J. Am. Chem. Soc. 2000, 122, 8168.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 14, 2012
3755