kilomentor

A topic that was much discussed at the recent Current Progress in Process Chemistry meeting in Princeton June 13-14th was enzyme catalysis of pharmaceutical process steps. Kilomentor has written about this subject before:

http://kilomentor.chemicalblogs.com/55_kilomentor/archive/775_biocatalytic_methods_of_enantioselective_synthesis_in_pharmaceutical_process_development.html

The essential message of this earlier blog article of mine was that for some common chemical transformations there are now so many different enzyme catalyst variants available that it is an odds-on bet that the chemist will be able to find one that can do such steps. These transformation comprise such things as ketone reductions, hydrogenation, P-450 type oxidations, dehydrations, esterifications, amidations, epoxidations, andnitrile hydrolysis.  This list is not exhaustive. At the June Current Process chemistry meeting,  Jacob Janey PhD, a Senior Research Chemist with Bristol Myers Squibb, gave a presentation titled Biocatalysts: Removal and Tracking during API Production. This talk addressed the disadvantage that, even though these reactions proceeded usefully, at the end of the reaction period when the protein catalyzing a reaction in water/ 60% organic solvent gets denatured, the mixture often turned the consistency of Jello making the isolation from the biocatalyst residue tedious. Dr. Janey reviewed the accepted methods for these isolations: centrifugation, evaporation followed by solid extraction, filtration, extraction and immobilization of the catalyst. He then listed the disadvantages of each method.

Centrifugation is typically only appropriate for small scale work because of the equipment. Also it is slow and decanting the top phase can be tedious.

The evaporation followed by solid extraction is performed by diluting the mixture of water and organic solvent with either acetonitrile or IPA and repeated evaporating to remove all the water as an azeotrope, then concentrating down to an organic slurry, filtering, and passing the organic layer through silica gel. The disadvantages here are that once again it is limited to small scale, it can be slow since the processing is long, and the filtration can still be a problem.

The filtration method is applicable to reaction mixtures with 2-50% organic cosolvent with water. First 200-400% Solka-Floc®   (powdered highly purified functional cellulose ) is added, then the mixture is acidify to pH below 3, then agitated and heated to about 50 C to denature the enzyme. The resultant is an aqueous precipitate  that can be filtered through cloth and rinsed through with 0.1 N HCl. This method has the disadvantage that it only works with aqueous acid soluble substrates (amines). It can also be a very slow filtration and the enzyme residue may break through into the filtrate to some extent. It is not be general for all enzymes. The advantage is that it is applicable on scale when it is applicable at all.

However, this talk was about the advantages of the extraction procedure and how it could be usefully adapted.

The method applies to an enzymatic transformation that has been conducted in water with 2-50% of an organic cosolvent such as methanol, DMSO or isopropyl alcohol (IPA). First, the pH is adjusted (if necessary) so that the substrate will be extractable into an essentially organic layer when it is introduced. Then, 0.5 relative volumes of alcohol (IPA or MeOH) is added. Then 0.5 relative volumes of isopropyl acetate or methyl t-butyl ether is added. The order of addition of these liquids is noted to be crucial. Then moderate agitation is employed to cause phase mixing to even out the concentration of liquids in between the  physical layers and then allow phase separation. The goal here is to achieve a 1:1:1 v/v/v mixture of water, alcohol and aprotic organic before the two layers separate. The bottom, more essentially organic, layer is then backwashed with 0.5 relative volumes of 30:70 v/v  alcohol/aprotic organic. The upper, more essentially organic wash is combined with the first upper, essentially organic layer and after being combined together they are in turn washed with water. The product is to be found in this wet combined alcohol/aprotic organic layer. The advantages of this method claimed are that it is scaleable, gives reliable easily visible quick separating cuts, and is reliable and adjustable to the particular process by adding more or less alcohol as needed to give best separation of the layers. I would judge this makes it the most general and scalable isolation procedure available. The admitted disadvantages are that significant product losses arise for relatively more polar substrate molecules. These are held more strongly in the water rich layer and so are lost to a greater degree. The method is of necessity more volume intensive and can be more lengthy since more additions are made to achieve superior layer separation.

The method seems to this author to be based on a method of separating an organic from an aqueous layer in the presence of insoluble materials that was patented by Reuel Shinnar and Roberto Mauri for The Research Foundation of the City University of New York in US 5,628,906. The patent authors say that “[T]he inherent advantage of this method is that it works effectively even in the presence of substances (solid or dispersed) that cause the formation of emulsions or stable dispersions. Though the effect of adding a modifier on the solubility of a solvent mixture is known, that such addition can result in rapid phase separation even in the presence of emulsion forming impurities has not heretofore been observed in connection with an extraction process”. I provide this reference to fully inform my readers. I have no way of knowing whether Dr. Janey  knew of this intellectual property or whether Bristol Myers Squib consider it relevant to this method for separating denatured enzyme from the reaction mixture.