Article
Practical Asymmetric Synthesis of (+)-erythro Mefloquine
Hydrochloride
William P. Hems,† William P. Jackson,† Peter Nightingale,‡ and Rob Bryant*
†Creative Chemistry, I-64 Building 1, Brunel Science Park, Kingston Lane, Uxbridge UB8 3PQ, U.K.
‡Development Chemicals Limited, UK Operations, 34 The Drive, Orpington, Kent BR6 9AP, U.K.
ABSTRACT: A highly enantioselective and cost efficient process for the synthesis of (+)-erythro mefloquine has been
developed. The key step is an enantioselective reduction of pyridyl ketone KI using transfer hydrogenation with formic acid as
the hydrogen source. The ratio of formic acid to NEt3 was found to be very important to achieving a highly efficient process.
efloquine is a highly efficacious antimalaria drug that is
approved in its racemic form for both treatment (in
asymmetric hydrogenation of KI using (S)XylBinap-RuCl2-
(S)Daipen, which gave the chiral alcohol in 88% ee, 92% yield
using 1 mol % of catalyst.6 These two reports provided
evidence for the asymmetric reduction approach we favored,
but we were worried by the economic issues of using both these
expensive Rh and Ru catalysts. Another outstanding catalytic
asymmetric reduction technology has been developed in recent
years for the enantioselective reduction of ketones. Mono-
sulfonated diamine Ru arene complexes, which utilize hydrogen
donors such as 2-propanol or formic acid, have been used to
reduce ketones with excellent enantioselectivities.7 In light of
the Merck result, we investigated whether transfer hydro-
genation (TH) would yield an efficient enantioselective
reduction process.
The pyridyl ketone KI was synthesized using known
literature procedures.8 When standard catalytic asymmetric
transfer hydrogenation conditions, using [(S,S)TsDpen]Ru(p-
cymene)Cl as the catalyst precursor were applied, the ketone
was reduced to the chiral alcohol in full conversion and in 96%
ee (Scheme 2). It was extremely gratifying to achieve such high
enantioselectivity, and so we continued to investigate the scope
of the reaction.
Amazingly, a screen of different solvents and catalysts led to
virtually no change in the enantioselectivity of the reaction. It
was pleasing to note that the use of TsDACH ligand9 in place
of TsDPEN resulted in equally excellent ee (96%), as this
added a cost benefit in terms of lower catalyst cost contribution
to the process.10 The parameter which had the most impact on
the reaction was the ratio of formic acid to NEt3. The results of
using different ratios of formic acid/NEt3 are summarized in
Table 1. We found that a narrow window existed in which the
reaction proceeded with exceptionally high TOF11 and
enantioselectivity. The reaction appeared fastest when the
molar ratio of formic acid to NEt3 was 1. Increasing the ratio
led to a much reduced reaction rate, but maintenance of the
enantioselectivity. Lowering the ratio to below 1 resulted in not
only a lower rate but also a lowering in ee. The optimized
conditions (s/c 1000,11 formic acid/NEt3 1:1, DMF, rt, 20 h)
were reproduced on 2 × 11 g input of pyridyl ketone KI, and
M
combination with artesunate) and prophylaxis (as monother-
apy) of malaria.1 The molecule contains two contiguous chiral
centres, which means that it can exist in four possible
stereoisomers (Figure 1).2 Larium, the Roche branded
formulation, contains rac-( )-erythro mefloquine HCl with
undetectable levels of the rac-threo stereoisomers. Although
rac-erythro mefloquine is an attractive antimalaria therapy, its
clinical utility has been compromised by CNS side effects. It
has been felt on theoretical grounds that these side effects
might be mainly a byproduct of the (−)-erythro stereoisomer.
In June 2009, Treague Ltd., a small biotech company based in
Cambridge, U.K., and Medicine for Malaria Venture (MMV3),
a specialist malaria funding charity, initiated phase 1 stage
clinical development of the single enantiomer (R,S)-erythro
mefloquine HCl for the treatment of malaria.
As part of that program, Development Chemicals Ltd./
Creative Chemistry were invited to identify manufacturing
options and to develop a process to provide single enantiomer
active ingredient for further clinical trials. The perceived benefit
of (+)-erythro mefloquine would be the lowering of the CNS
side effects and, therefore, better compliance. A higher dose,
however, makes the development even more sensitive to cost of
goods in that it is not desirable for the single enantiomer to be
of greater cost than the racemate. With the fact that the API
would be destined for use in combination therapies to be
dispensed in predominantly disadvantaged regions of the world,
MMV and Treague set a very tough target of half the cost of the
current racemate! Therefore, it was extremely important that a
cost-effective asymmetric synthesis of (+)-erythro mefloquine
was developed. Although the resolution of the rac-erythro
mefloquine has been reported, this route was discounted as
inevitably the route that produced material of at least twice the
cost of the racemate.4 It was interesting to note, however, that
the synthesis of rac mefloquine went via pyridyl ketone KI
(Scheme 1).
Roche has reported the Rh(bisphosphine) hydrogenation of
pyridyl ketone KI. The product carbinol was typically in the
range 82−92% ee in 84−92% yield.5 After asymmetric
reduction of the carbonyl moiety, the pyridyl ring was
hydrogenated using a heterogeneous Pt catalyst without
isomerisation. More recently, Merck has also published the
Received: December 5, 2011
Published: February 21, 2012
© 2012 American Chemical Society
461
dx.doi.org/10.1021/op200354f | Org. Process Res. Dev. 2012, 16, 461−463