Access to large amounts of core building blocks is essential in drug discovery programs. It enables rapid synthesis of a series which can then give structure activity relationships (SAR) of the drugs and their target. SAR data allows medicinal chemists to determine the group(s) in a structure evoke a change in activity, whether positive or negative. A comprehensive SAR directs medical chemists towards the next target and greatly enhances the drug discover effort.
Figure 1: Quinoline and furopyridine cores.
When I joined this project there were two main pharmacophores that were being explored, quinolines and furopyridines (Figure 1). The quinoline core has been utilized for decades in drugs, one notable example being chloroquine, historically one of the most successful anti-malarial drugs to be marketed. The quinoline core is commercially available with various handles/substitution patterns for derivatization. The furopyridine core if much less common in clinical drugs. The most notable example of a furopyridine core in a drug is the HIV protease inhibitor L-754,394 (Figure 2).
Figure 2: L-754,394
The furopyridine core is less common in clinical drugs due in part to the lack of synthetic routes to the core. The routes that are published suffer from poor to moderate yields and give restricted opportunities to derivatize the core. Indeed, when I joined this project, progress in the furopyridine series was severely limited due to the inaccessibility of the core with the required handles. When the project was first initiated, a bromofuropyridine 1 was commercially available. This was subjected to Suzuki coupling followed by bromination to give another synthetic handle (Scheme 1). This route was expensive, slow and problematic, often affording side-products that were poly-brominated and difficult to separate by column chromatography.
Scheme 1: Initial route to furopyridine derivatives.
This route quickly became redundant when the starting material 1 was no longer commercially available. In any case an alternative, more effective, route was needed to optimize access to the core with the required handles for analogue synthesis.
An alternative route was derived from a synthesis published in 1986 by Hiroyuki Morita and Shunsaku Shiotani (Scheme 2). Although the route looks good on paper it has some severe drawbacks. Firstly, the saponification-decarboxylation of ester 6 is extremely unreliable. The ester is surprising stable to acidic and basic conditions. When increasingly harsh conditions are used it often leads to decomposition of the core over hydrolysis. 41% was the best yield achieved in over 30 attempts using various conditions. Most frustratingly, when effective saponification conditions (tBuOH/aq. LiOH at reflux) were found, the acid was isolated, and the decarboxylation could not be forced under acidic, basic or thermal conditions. This suggests the basic saponification conditions inhibited the decarboxylation and acidic hydrolysis is required to enable the decarboxylation.
The second issue was the reactivity of the bromine and triflate of compound 8 to cross-coupling. Considerable effort was invested trying to find conditions for selectively coupling the triflate but in no case could complete selectivity be achieved.
Scheme 2: Unoptimized route to the furopyridine core.
To overcome these issues and streamline the route I decided to start with a dichloro derivative of nicotinic acid. Chlorine is much less active to cross-coupling than bromine and hence should allow for the triflate to couple exclusively. To address the saponification-decarboxylation step I decided to synthesize a tert-butyl ester which can be easily removed under mild acidic conditions. The updated route is shown in Scheme 3.
Scheme 3: Optimized route to the furopyridine core.
Gratifyingly the route straightforward and reliable. Out of the four steps only 2 required purification via column chromatography. It is effective on a large scale and several grams of the furopyridine core can be made. Started from 5 grams of the dichloronicotinic acid 9, over 3 grams for furopyridine 12 was generated, amounting to a 70% yield over 3 steps.
Furthering this when compound 13 is subjected to palladium cross-coupling, the triflate shows much higher reactivity and exclusively couples first. This optimized route to the furopyridine core is now allowing us to generate a large set of compounds based on the furopyridine core. A large number have been sent for testing in an enzymatic assay for CaMKK2 with our collaborators and the results will be discussed in the future.
A robust synthetic route is key to a successful medicinal chemistry campaign. Nice work solving this problem for the CAMKK2 project.