In 2008 the Takeuchi group at Niigata University reported the total synthesis of cucurbitoside A, one of a series partially acylated phenolic glycosides. The cucurbitosides are found in the seeds of squashes and pumpkins of the genus Cucurbita. Water extracts of these seeds have shown hepatoprotective effects, although to date the causative agent for those effects is not known. Meanwhile other closely related furanosides have also demonstrated biological activities. The Takeuchi group, therefore, have targeted the synthesis of various members of the cucurbitosides in order to advance study of the these compounds and now report the synthesis of 8 cucurbitoside-like compounds. Four are naturally occuring cucurbitosides, two are closely related seguinosides, and two are unnatural analogs.
The researchers could have simply individually synthesized each of these based on their prior synthesis of cucurbitoside A. This, however, requires a large number of individual reactions. For eight compounds it’s not that bad, but as the project continues it is assumed that other analogs will be desired and with each one a separate total synthesis becomes increasingly tiresome. The researchers therefore chose to use a mixture synthesis approach to preparing the 8 compounds.
In a mixture synthesis multiple substrates are reacted in a single reaction vessel resulting in a mixture of multiple products. Each of the substrates is tagged with an unique tag. After several reactions the mixture is demixed using the tags which are then removed to provide each of the desired products. The advantage is that one does not have to make each compound separately thereby reducing the number of reactions that have to be run. The key, of course, is that you need a set of tags that can be used for the separation. Ideally the tag should have the following characteristics:
- Chemical inertness – allows multiple reaction conditions to be conducted
- Solution phase chemistry – provides better reaction kinetics and allows reaction monitoring
- Predictable separation order – simplifies separation and identification of mixture components
- Orthogonal separation mechanism – tolerates diversity within the mixture without adversely effecting separation
That’s fluorous tags to a T. Fluorous tagged compounds can separated by fluorous HPLC based on the length or number of fluorous tags present. So the longer the tag, the longer the retention time on the fluorous column. Several groups have used fluorous mixture synthesis in the production of chemical libraries. The orthogonality of the fluorous tags and separation also allows for double mixture strategies as exemplified by the work of Curran and Wilcox who used fluorous tags and oligoethylene glycol tags in tandem. The Takeuchi group decided to use a similar tack for their cucurbitoside synthesis using fluorous tags and acyl tags in tandem in a double mixture synthesis.
The researchers synthetic strategy is shown above. They prepared two different glycosides each tagged with a Rf6 benzyl or a Rf8 benzyl tag. Fluorous solid phase extraction (FSPE) was used throughout the synthesis of each of these two compounds. Four different furanose elements were prepared each with a different acyl tag. Three of the acyl tags were various substituted benzoyl group with the fourth was an alkyl acyl. The two fluorous tagged glycosides were mixed with the four furanosyl trichloroacetimidates to produce a mixture or eight (2 x 4) disaccharides. The mixture could then be demixed using FHPLC to provide two new mixtures each containing four compounds; all the Rf6 tagged compounds in one and all the Rf8 compounds in the other. These mixtures could then be separated by reverse phase HPLC based on the polarity of the acyl tags to provide all eight compounds.
In practice the researchers obtained the mixture of 8 compounds which they first pre-purified by FSPE. The demixing was not conducted using fluorous HPLC followed by two reverse phase HPLC’s, but rather conducted a single separation by serially connecting the two different columns. This reduces 3 HPLC purifications to a single one. The tricky part of doing that is that one has to use the same gradient through each column which obviously means that one or both separations are not optimized. Seen below are several chromatograms of the mixtures. A couple of things to note. First, one can see that the fluorous separation is primarily two peaks; the Rf6 compounds and the Rf8 compounds, although compounds 20b and 20f, the alkyl acyl tagged compounds were apart from the other compounds. This reflects a secondary separation mechanism separate from the fluorous separation, so total orthogonality is not observed. Next, note that baseline separation between 20c and 20h is not observed using RP-HPLC alone. The serial separation using both modes, however, does result in baseline separation of all compounds. After demixing all the tags are removed to provide the 8 cucurbitoside-like compounds.
As a mixture synthesis the preparation of these 8 compounds does not really save much effort since the mixing occurs at the end of the synthesis. The authors realize this and concede the point in their conclusion, but offer the synthesis as proof of concept for the double mixture synthesis and correctly point out that efficiencies are gained as more reaction steps are conducted after the mixing step.