Getting R(-)- Praziquantel by Resolution or Dynamic Resolution of an Intermediate

19 Sep
Published by kilomentor


In order to make full use of one's special personal creativity upon first exposure to a problem, I find it can be productive to look at a problem without knowing, in detail, the strategy that others have preferred.
The praziquantel problem statement could be expressed as:
 Prepare R-(-) praziquantel by a scaleable process that can realistically be optimized to pennies per gram. The deviation is that only the racemate is inexpensive so far. Another expression of the deviation from the present reality is that syntheses of the eutomer are too expensive.
The enantiomer that is characterized by most of the pharmaceutical activity is called the eutomer. For praziquantel, the R-(-) compound is the eutomer. The enantiomer that carries little or no activity and often introduces side effects, such as bad taste in the case of praziquantel, is called the distomer. I will use this terminology repeatedly.
The options for a successful outcome here seem to be restricted to a  selection among (1) highly enantioselective synthesis, (2) dynamic resolution, or (3) separation of enantiomers and conversion of each separately to eutomer; in order to be cost effective, if the synthesis is not enantiospecific, both enantiomers of Praziquantel need to convert into the eutomer.
Praziquantel is quite a flat molecule. In the tricyclic ring system,  9 of 14 atoms are approximately sp2 type and the substituent that is being added at the chiral centre [11b by standard numbering]  is just a hydrogen atom. Getting a significant enantiotopic preference from a reagent or catalyst is I expect going to be very challenging. I would not expect the optimal set of conditions to be rugged. That is because the enantiotopicity needs to be imported from a reagent or catalyst to a substrate where the two faces will not appear much different the difference in the free energy of activation for making the eutomer precursor and the distomer precursor is likely to be small and hence sensitive to quite a few of the reaction condition parameters. If a success were to be achieved on alaboratory scale, the victory could easily vanish on scaling up because heat transfer, mass transfer and stirring only become more difficult. For this reason, my inclination is to look elsewhere for a scaleable inexpensive solution.
2-[N-cyclohexylcarbonyl(aminomethyl)]1,2,3,4-tetrahydro-isoquinoline is an intermediate reported by H. Rupe and W. Frey, Helv. Chim. Acta. 22, 673 (1939). It is synthesized from isoquinoline, cyanide, hydrogen and hexahydrobenzoyl chloride in two steps. Cyanide and cyclohexylcabonyl radicals are added 1-2 to the isoquinoline and then the cyanide reduced to a primary amine to which the acyl group migrates. The product is a racemic, basic, secondary amine that looks like it can be involved in intramolecular bidentate metal complexing. The potential chiral centre is adjacent to the basic site. This is the most preferred situation for enantiomeric separation. One can imagine the benzylic hydrogen on carbon adjacent to the secondary amine equilibrating under conditions of strong base (NaH) in the presence of a catalytic hydride acceptor such as fluorenone. An alternative to this is the efficient and mild ruthenium-catalyzed racemisation of amines as applied to the synthesis of enantiomerically pure amines, Tetrahedron Lett., 43, 4699-4702.  
If the R-2-[N-cyclohexylcarbonyl(aminomethyl)]1,2,3,4-tetrahydro-isoquinoline can be obtained the synthesis is completed by reaction with chloroacetyl chloride or an equivalent.
Lets stop my consideration for now. My sense is for this system to work properly some time for percolation in other minds stocked with other insights is required.


MatTodd's picture

Yes indeed these are good ideas. Resolution is something we've recently solved for a different intermediate that is derivable from PZQ itself, but resolution of a different PZQ precursor is certainly possible. The racemisation is possible through a number of means and the paper you mention is one approach. I wonder if there are others on substrates directly related to those we might want to use here? A DKR would be an efficient solution, of course, and relies on having the two competing functionalities happy to coexist in one pot. If you're interested in doing more, are there any literature examples of systems that could be applied to our case, to narrow down the possibilities? You could start a new page on this maybe as a child page to the "starting from the racemate" page.

This piece of work has been done by us as part of a brief WHO funded investigation into processes to PZQ.
Full details of this promising approach will be posted here once approval has been obtained from WHO.

MatTodd's picture

Great - very much looking forward to this!

We have prepared a summary for TSL that we will forward to you for insertion. We sent it to WHO yesterday for approval.
Regards, CC