Regioselective Semisynthetic Modification of a Cardiotonic Steroid Glycoside

Semisynthetic derivatization of complex natural products can often be an effective medicinal chemistry strategy to improve upon the potency/selectivity profile or pharmacokinetic properties of a naturally occurring bioactive molecule. For example, the potent and broad-spectrum antifungal drug CANCIDAS is a synthetic derivative of the echinocandin macrocyclic lipopeptide natural product, pneumocandin Bo. A significant challenge to this type of approach is the development and implementation of highly selective synthetic methods that can effectively manipulate a single part of a molecule in the presence of other functionality with similar reactivity. Along these lines, the regioselective acylation of polyol natural products is an active area of academic research that could eventually impact the very nature of the drug development process by rendering previously inaccessible semisynthetic compounds readily available. Takeo Kawabata's research group at Kyoto University has recently described the regioselective acylation of the cardiotonic steroid glycoside, lanatoside C, which is composed of the aglycone digoxigenin, linked via the steroidal C3-oxygen to a complex tetrasaccharide (structure shown above). Both the closely related steroid digoxin (1), as well as lanatoside C (2), are clinically used cardiotonic chemical entities. The Kawabata group has developed conditions to selectively monoacylate either the C(3'''')-OH or the C(4'''')-OH of the terminal glucopyranoside of the western tetrasaccharide of 2.

N,N-dimethylaminopyridine (DMAP)-catalyzed acylation of lanatoside C gives the C(3'''')-O-acylate (3) with 97% regioselectivity due to the inherent nucleophilic reactivity of the substrate. On the other hand, acylation of lanatoside C in the presence of the complex organocatalyst (A) affords the C(4'''')-O-acylated product with about 90% catalyst-controlled regioselectivity. The authors provide a very interesting hypothetical transition state assembly model to account for the observed outcome. In this picture, the catalyst-substrate interaction involves hydrogen bonds between (C6'''')-OH (to a carbonyl) and C(3'''')-oxygen (to indole N-H) and the catalyst. A molecular modeling conformational analysis of lanatoside C was also conducted. It is quite remarkable that Kawabata's organocatalyst (A) can distinguish between the no less than eight free hydroxy groups within 2 to provide essentially a single acylation product (4), an outcome that is completely divergent from the inherent reactivity patterns of the natural product. This outstanding effort is reminiscent of the now classic pioneering work done by Scott Miller's group on erythromycin A, apoptolidin A and vancomycin.