Nucleosides have emerged as a key chemical class in successful antiviral and anticancer drugs with nearly half of currently marketed drugs possessing these cores.1 However, the biologically active nucleoside phosphates, which are generated in vivo by phosphorylation, are poor drug candidates due to issues of permeability and stability. One pronucleotide strategy developed to address these challenges has been described by McGuigan. His ProTide platform introduces a 5’-arlyoxy phosphoramidite to the drug candidate which can result in improved cell permeability and rate of phosphorylation compared with non-phosphoramidite containing nucleosides.2,3
However the introduction of the phosphoramadite in the prodrug can have a significant effect on the potency, toxicity and rate of metabolism, effects which can be associated with the stereochemistry at the phosphoramidite phosphorus. Unlike for carbon chemistry, where stereocontrol is a sophisticated branch of asymmetric catalysis, p-chiral chemistry is far less developed. Although methods do exist to steer products towards a preferred p-chiral isomer, such as dynamic kinetic asymmetric transformation (DYKAT) using chiral auxiliaries or desymmetrisation of achiral species,3 these approaches have significant drawbacks with the former suffering from poor selectivities and low catalytic turnovers, whilst the latter is often a complicated multistep synthesis.
In addition to the stereo-isomeric challenges during prodrug assembly there also remains the challenge of chemoselectivity for 5’ versus 3’ phosphoramidation. With these challenges in mind, I wanted to highlight an excellent paper published by a team from the process research and development group of Merck & Co., USA, where they report the first good example of a catalyst being designed to address the issue of stereo- and chemoselectivity in the synthesis of pronucleotide prodrug candidates. In their paper they focused on a hepatitis C virus RNA polymerase inhibitor currently in late stage clinical trials (MK-3682, Figure 1).4
Figure 1: General phosphoramidation scheme and the best in-class catalysts developed to effect the coupling. Yield is the total yield of phosphoramidite isolated, chemoselectivity for 5’ vs 3’ is represented by ratio 5’:3’and d.r is the ratio of P(R) to P(S).
Using mechanistic studies, computational modeling and an understanding about the enzymatic mechanism of P-O bond formation in the phosphorylation of nucleosides, the team successfully developed several small molecule organic catalysts that mimic the concomitant series of activation modes used by enzymes to effect the P-O bond formation. Early studies identified carbamates as a privileged class for controlling stereo- and chemoselectivities with catalyst (R)-B being the best of the first generation catalysts developed. Using computational modeling the carbamate was theorized to carryout 3 roles; leaving group activation, general base catalysis and oxyanion stabilization via a pentavalent transition state, giving rise to a 2.3kcal/mol differentiation between the desired R-stereochemistry and the S-stereochemistry at the phosphoramidite phosphorus (Figure 2). Catalyst I, the best of the catalysts reported, evolved from a conscious effort to increase the transition state differentiation by decreasing the entropy of the system via linkage of catalyst (R)-B. As highlighted in Figure 1 the linked catalyst, Catalyst I, was able to achieve excellent yields and high selectivities for both the desired 5’ product (99:1 in favour of the 5’ product) and with a d.r (of 99:1) in favour of the desired P(R) isomer.
Figure 2: Transition state model showing multiple catalyst modes of action. Reproduced from reference 4.
This work demonstrates an excellent step forward in the controlled synthesis of pronucleotide prodrugs by continuing to employ rational design beyond the discovery phase SAR, well into the late stage development of the prodrug. Moreover the published work is an elegant example of the power of using an interdisciplinary approach to solve chemical problems via a rational design cycle.
Written By Jason A. Gillespie
- P. Jordheim, D. Durantel, F. Zoulim, C. Dumontet, Nat. Rev. Drug Discov. 12, 447–464 (2013).
- Cahard, C. McGuigan, J. Balzarini, Mini Rev. Med. Chem. 4, 371–381 (2004).
- J. Sofia et al., J. Med. Chem. 53, 7202–7218 (2010).
- A. DiRocco et al., Science, 356, 426–430 (2017).